Ion-chromatogram libraries assembly in DIA proteomic analysis of post-exercise skeletal muscle in prediabetic subjects

Ion-chromatogram libraries assembly in DIA proteomic analysis of post-exercise skeletal muscle in prediabetic subjects | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Article Ion-chromatogram libraries assembly in DIA proteomic analysis of post-exercise skeletal muscle in prediabetic subjects Anna Czajkowska, Łukasz Szczerbiński, Macin Czajkowski, Anna Citko-Rojewska, and 4 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-6331082/v1 This work is licensed under a CC BY 4.0 License Status: Under Review Version 1 posted 10 You are reading this latest preprint version Abstract Physical exercise of even a moderate intensity is beneficial in both the prevention of prediabetes and management of Type 2 diabetes mellitus, as skeletal muscle is a primary tissue responsible for glucose uptake. Exercise-evoked proteomic alterations in muscle of subjects with prediabetes are of great importance for the study of relationships between insulin resistance and exercise. Due to its molecular composition proteomic analysis of skeletal muscle is challenging. To identify optimum approach, we compared various ion-chromatogram libraries assembled with the use of off-line high-pH fractionation (HpH), gas-phase fractionation (GPF) and libraryless DirectDIA™ in LC/MS/HRMS DIA proteomic analysis of muscle from normoglycemic (NGT) and prediabetic (IGT) subjects after 3 months of supervised, mixed-mode exercise. GPF-fractionated, hybrid DDA/DIA libraries yielded the best overall balance between the speed of preparation, data collection and protein identification. Analysis revealed, that despite 3-month exercise intervention skeletal muscle from IGT subjects displayed significant alterations in pathways and molecules relevant to muscle contraction, extracellular matrix composition and protein synthesis as compared to NGT counterparts. In conclusion, our study underlines the importance of the selection of appropriate approach in the analysis of challenging clinical samples and reveals the potential explanation for deficiency of muscle function in the prediabetic state. Biological sciences/Biochemistry/Proteomics Health sciences/Endocrinology/Endocrine system and metabolic diseases Biological sciences/Biochemistry/Hormones Health sciences/Diseases/Endocrine system and metabolic diseases/Diabetes/Type 2 diabetes mellitus Health sciences/Diseases/Endocrine system and metabolic diseases/Metabolic syndrome Health sciences/Diseases/Endocrine system and metabolic diseases/Obesity Health sciences/Diseases/Endocrine system and metabolic diseases/Pre diabetes Health sciences/Anatomy/Musculoskeletal system/Muscle/Skeletal muscle Health sciences/Biomarkers/Predictive markers Biological sciences/Physiology/Metabolism/Metabolic diseases/Diabetes/Type 2 diabetes mellitus Physical sciences/Chemistry/Analytical chemistry/Mass spectrometry Physical sciences/Chemistry/Analytical chemistry/Medical and clinical diagnostics Biological sciences/Systems biology/Biochemical networks Biological sciences/Molecular biology/Proteomics Figures Figure 1 Figure 2 Figure 3 Figure 4 Figure 5 Figure 6 Figure 7 1. Introduction Type 2 diabetes mellitus (T2DM) and its comorbidities such as cardiovascular-related disorders, hyperlipidemia, obesity, nephropathy, neuropathy and decrease in cognitive function and dementia become a major health-related burden worldwide 1,2 . Progression towards T2DM is connected with increasing insulin resistance of major insulin-sensitive tissues (muscle, adipose tissue, liver) with subsequent induction of pre-diabetic state. This usually asymptomatic, obscure metabolic abnormality can manifest itself as isolated impaired fasting glucose (IFG – fasting blood glucose between 100mg/dl and 125mg/dl) or impaired glucose tolerance (IGT – 2 hour blood glucose after 75g oral glucose challenge between 140mg/dl and 199mg/dl) and elevated blood insulin 3 . As prediabetes is often overlooked, pre-diabetes progresses toward symptomatic T2DM after b-cell failure, when pancreatic islets are unable to compensate for increasing systemic insulin demands. This crucial watershed event marks an irreversible stage of T2DM pathogenesis 4,5 . Currently, due to development of various pharmacotherapies (SGLT2 inhibitors, DPP4 inhibitors, GLP1 analogs and receptor agonists, insulin sensitizers and long-acting insulin analogs) case-tailored T2DM management is increasingly successful, leading to both a decrease in hyperglycemia and T2DM comorbidities 6,7 . Contrary to fully developed T2DM, b-cell dysfunction and systemic insulin resistance in prediabetes is reversible and can be successfully managed by approaches such as bariatric surgery, pharmacotherapy, lifestyle changes, physical exercise and their combinations 8-11 . From the above, physical exercise appears as an easily accessible, universal remedy in both the prediabetes prevention, augmentation of its reversal or inhibition of its progression towards T2DM 10,12,13 . Moreover, physical exercise has shown beneficial effects in the prevention of other metabolic-syndrome connected comorbidities such as cognitive impairment and dementia 14-16 . Metabolic health of muscle tissue is of great importance for both the physical fitness and whole-body energy metabolism. As a major tissue in both insulin-stimulated and insulin-independent (exercise-evoked) uptake and metabolism of plasma glucose, its role in prediabetes prevention and its reversal cannot be overlooked. Prediabetes was shown to induce detrimental effects in both the metabolic and contractile muscle function 17-20 . Crucially, exercise in different forms and modalities was shown to improve both metabolic and functional abnormalities observed in prediabetes 21,22 . Taking all of the above into account, analysis of the impact of exercise on the proteome of prediabetic skeletal muscle is crucial for the elucidation of its beneficial effects. Moreover, the subject of equal importance is the identification of persistent negative effects of prediabetes, which are not fully corrected by the physical activity. However, proteomic analysis of skeletal muscle is burdensome due to several difficulties arising from the biological and molecular characteristics. At the macroscopic level these include site-dependent variations in connective tissue content, extracellular matrix deposition and adipose tissue distribution, which – despite careful sampling and post-biopsy cleanup – induce significant variation in sample composition. Moreover, developed striated muscle is composed mainly of chains of biological polymers involved in muscle contraction, such as myosin, actin, tropomyosin and titin chains 23,24 . This creates so-called “tip of the iceberg effect”, similar to that of plasma proteome, where a number of proteins are greatly overrepresented in non-depleted samples, which subsequently masks low-abundant proteins 25,26 . Thus, methods tailored for proteomic analysis of skeletal muscle cannot be directly derived from those developed using easily accessible matrices, such as cell-culture material. Cultured cells are inherently less demanding and yield high proteome coverage levels, presenting best-case scenarios in proteomic sample analysis. Robust methods for proteomic analysis of clinical skeletal muscle samples need to address the specificity of this tissue at both the LC/MS/MS level and subsequent bioinformatic analysis. At the HPLC peptide separation level, it requires the use of traps and columns resistant to clogging by long biological polymers, flow-reversible for easy cleanup and restoration of its initial performance, with robust retention time stability after hundreds of samples 27,28 . Ideally, with low operating pressure and long separation length and adequate resolution to address “tip-of the iceberg” effect of non-depleted samples. At the MS level, it dictates the use of data-independent acquisition (DIA), ideally supplemented with retention time calibration peptides, which is better suited for clinical applications due to its reproducibility and non-stochastic nature of identification, unlike data-dependent acquisition (DDA) 29,30 . Finally, employment of DIA requires assembly of appropriate ion-chromatogram libraries, or selection of libraryless, pure bioinformatic identification and quantitation pipeline 31-33 . To address above challenges, we constructed a pipeline tailored for proteomic analysis of clinical skeletal muscle biopsies. We employed mPAC semiconductor-technology chromatography columns and traps, retention-time calibration peptides and staggered-windows DIA analysis to address reproducibility issues inherent to challenging sample matrices. As employment of different approaches to ion-chromatogram library creation can lead to distinct biological findings, we compared various ion-chromatogram libraries based on off-line high-pH fractionation (HpH), gas-phase fractionation (GPF) and libraryless DirectDIA TM approach. Finally, we performed proteomic analysis of muscle from normoglycemic (NGT) and prediabetic (IGT) subjects after 3 months of supervised, mixed-mode exercise 2. Materials and Methods 2.1 Study participants and sample collection: The muscle biopsy samples were collected from participants of the “Bialystok Exercise Study in Diabetes (BESD)”, conducted by the Department of Endocrinology, Diabetology and Internal Medicine and Clinical Research Centre of the Medical University of Bialystok 22,34 . The study and its procedures were accepted by Ethics Committee of the Medical University of Bialystok (No. R-I-002/469/2014). Each participant signed informed consent before participation and subsequent sample collection. Briefly, the study cohort involved physically inactive men with various degrees of dysglycemia, with a BMI of 25-35 kg/m 2 and leading a sedentary lifestyle. All participants underwent three-month exercise program that included supervised training sessions at a local fitness center. Studied patients participated in a 3-month exercise intervention consisting of mixed training, aerobic and strength exercises, performed for 85 minutes 3 times a week, totaling in 36 sessions over 12 weeks 22,34 . The routine started with a 15-minute warm-up, followed by 40 minutes of strength exercise of major muscle groups (60-75% of single repetition maximum, adjusted weekly) and 30 minutes of moderate-intensity endurance exercise (60–70% of individual VO 2 max), all supervised by an exercise technician. The myWellness system (Technogym, Cesena, Italy) recorded each session for consistency. Dietary supervision aimed to eliminate diet-related variables. For the assessment of proposed methodology, we selected only the post-exercise bioptates from non-diabetic, normoglycemic, with normal glucose tolerance (NGT group; n=13, fasting plasma glucose (FPG) <100 mg/dL and 2-h glucose during OGTT (2h-OGTT-GLU) <140 mg/dL) and prediabetic, impaired glucose tolerance group (IGT group; n=11; FPG <100 mg/dL, 2h-OGTT-GLU 140–199 mg/dL). Analysis was performed on 2 independent bioptates from each participant, yielding a total of 26 NGT and 22 IGT samples (total of 48 individual samples). Post-exercise anthropometric characteristics of NGT and IGT groups are presented in Supplement 1 (Table S1) .The vastus lateralis muscle (VL) biopsies were obtained from fasted participants 48 hours after the last exercise session with the use of a percutaneous suction needle 35 . Excess blood, connective tissue and fat was removed after visual inspection immediately after collection. All samples were snap-frozen and stored in LN 2 until further analysis. 2.2 Experimental Design: Experimental design of the study is presented in Figure 1 . 2.3 Sample preparation: Protein extraction and digestion was performed with the use of sodium deoxycholate-assisted sample preparation with phase transfer SDC extraction and lipid depletion technique according to Leon et al. 36 . Briefly, approx. 25-30mg of lateral thigh muscle biopsy was pulverized in LN 2 , suspended in tissue lysis buffer (1:10 w/v; 50mM ammonium bicarbonate (ABC), 5% sodium deoxycholate (SDC), 5mM TCEP) and sonicated for 2 x 30 sec. on ice at 50% power (Sonics&Materials VCX-130 Ultrasonic Processor). Subsequently, samples were denaturated by heating for 30 min. at 60°C, alkylated with iodoacetamide (IAA) for 30 min. at room temperature (IAA, 1M in ABC buffer, 15mM final concentration) and centrifuged for 5min. at 14000g to sediment tissue debris. Protein concentration was measured with BCA assay (reducing-agent compatible, Thermo Scientific) and protein content was normalized to 2.5 μg/μl with 5% SDC in 50mM ABC buffer. Digestion-ready samples were stored in-80°C until further analysis. For sample digestion, 10μl of sample equivalent to 25μg of protein, was diluted to 90μl with 50mM ABC. Trypsin/LysC mix (2.5μl from 0.4μg/μl stock in 50mM acetic acid) was added at a ratio of 1:25 enzyme/protein (1:50 ratio for each of the enzymes). Samples were incubated at 37°C for 12h with shaking (600 rpm) in Thermoblock, and subsequently cooled and kept at 4°C until next working day. After addition of TFA to 0.5% concentration and brief vortexing, precipitated SDC and lipids were extracted 3 times with 100μl of ethyl acetate. Each time samples were vortexed and centrifuged for 5 min at 14000g. Residual ethyl acetate was evaporated by heating the open tubes at 60°C for 30 minutes with gentle shaking. Subsequently, 10 injection equivalents of iRT peptides (Biognosys) were added to each sample. Additionally, approx. 20μg of protein was taken from each homogenized biopsy to create study-wide mixed sample. After protein digestion, this mixed sample was used for offline high-pH fractionation (HpH), gas-phase fractionation (GPF) and for repetitive injections performed at the library selection stage. All injection-ready samples were stored in -80°C until further analysis. 2.4 LC/MS/MS analysis: The same optimized chromatographic conditions were used for both data-dependent (DDA) and data-independent (DIA) analysis. Detailed description and results of optimization of chromatographic gradient, column flow and sample loading for µPAC™ pillar array trap and column are presented in Supplement 1 (Table S2-S4). All optimization steps were performed on HeLa digest standard (Thermo Scientific) run in triplicate in DDA mode. LC/MS/MS analysis was performed on Q-Exactive MS equipped with IonFlex II source and silica glass emitter (10μm or 20μm I.D. PicoTip, flow-rate dependent; NewObjective) working at 1.75kV in +ESI mode. Thermo Scientific 3500RSLC nanoLC configured for trap-elute was used in all chromatographic separations. For standard runs, samples were loaded at 750ng of peptides onto µPAC™ nano pillar array trap at 10μl/min (2% ACN, 0.2% TFA) and resolved on 50cm µPAC™ pillar array column (channel length 500mm, channel width 315μm, pillar height 18μm, pillar diameter 5μm, interpillar distance 2.5μm) at 300nl/min using 115min multi step reverse-phase binary gradient (A – 0.2% FA in H 2 O, LC/MS; B – 0.2% FA in 90% acetonitrile, LC/MS; 5%B-4min, 30%B-71min, 45%B-94min, 90%B-95min, 90%B-103min, 5%B-104min, 5%B-115min). Total sample run time excluding autosampler draw equaled 120min. Daily routine started with mass calibration, nanoLC solvent purge and column equilibration. Each batch of 11 samples was preceded by iRT-spiked QC sample (300ng of HeLa digest standard, Thermo Scientific) analyzed with 55 min gradient and DDA acquisition. Study-wide analysis of sample-spiked iRT peptide retention time variation, FWHM and mass accuracy and total ion chromatograms of all experimental samples and respective mass accuracy at MS and MS/MS level is presented in Supplement 1 (Figure S1 and S2) . 2.5 Data-dependent (DDA) acquisition: DDA analysis of eluting peptides for both the optimization of chromatographic conditions and library generation was performed using Top 20 method with following settings: MS1 resolution of 70k, ACG target 3e 6 , max. IT 50ms, scan range 385-1015m/z; MS2 resolution 17k, ACG target 2e 5 , max. IT 55ms, isolation 2m/z with 0.4mz offset to include most of isotopologues in case of multiply-charged peptides. Top 20 peptides (excluding isotopes) were selected for fragmentation with normalized collision energy (NCE) lowered to 27, if following criteria were fulfilled: charged state 2-5, preferred peptide match, min. ACG 6e 3 and intensity threshold 1.1e 5 , with dynamic exclusion set at 15sec. 2.6 Data-independent (DIA) acquisition and gas-phase fractionation (GPF): DIA method for library generation and sample analysis (referred as 24m/z STW in the manuscript) was designed on the basis of staggered windows approach by Searle, Pino and Amodei 31,37-39 . Method consisted of 52 staggered, 24m/z wide windows (12m/z overlap), covering 400-1000 m/z (referred as 24SW method) with MS1 scan covering 385-1015m/z performed after 26 windows (Supplement 1, Figure S3; Supplement 2, Table S1) . Mass range was selected on the basis of distribution of peptide m/z values from DDA run of mixed muscle sample. Window edges were placed in peptide forbidden zones to account for non-rectangular characteristics of Q-Exactive quadrupole mass isolation. Window placement was designed with Skyline. MS1 resolution was set at 35k, ACG at 1e 6 and max. IT of 60ms. Windowed MS2 was performed at 17k resolution, with ACG at 1e 6 , max. IT of 60ms, MSX count of 1 with isochronous IT set to ON and 27 NCE at default charge state of 3. For gas-phase fractionation (GPF) with staggered windows approach (4 m/z GPF-STW), 750ng of peptides was resolved for each of the 6 gas-phase fractions. Gas-phase fractionation consisted of 6 overlapping segments 102m/z wide (2m/z overlap) jointly covering 400-1000 m/z mass range (Supplement 2, Table S2) . Each gas-phase fraction consisted of 52 staggered, 4 m/z wide windows (2m/z overlap), with MS1 scan covering respective mas range performed after 26 windows and utilized identical MS1 and MS2 settings as DIA method described above. Results of GPF with staggered windows presented in Supplement 1 (Figure S3). 2.7 Off-line high-pH peptide fractionation with fraction concatenation: For HpH fractionation with fraction concatenation (HpH), post-digestion peptides from mixed sample were evaporated to dryness using Labconco Vacuum Concentrator at 45°C under 0.05atm residual pressure. Subsequently, peptides were dissolved at 5μg/μl in 2% ACN in 100mM NH 4 OH with 0.001% Zwittergent 3-16 through vortexing and brief sonication. Up to 100μg (20μl) of peptides were loaded onto C18 Waters X-Bridge Peptide BEH Column (300Å, 3.5 µm, 1 mm X 150 mm) and resolved using Dionex 3500RSLC. Peptides were eluted at a flow rate of 100μl/min using a 45 min, reverse-phase binary gradient (A -2% ACN in 10mM NH 4 OH, B – 90% ACN in 10mM NH 4 OH; gradient 3.3%B for 2.5min, 25%B at 32.5min, 40%B at 35min, 90%B at 35.5min for 3.5min, 3.3%B at 42.5min for 2.5min) under UV monitoring at 214nm (peptide bond) and 280nm (tyrosine, tryptophan). Peptide fractions eluting for up to 42 min were collected every 140 seconds into 96 well 500µl LoBind plate (Eppendorf). Resulting 18 fractions were pooled to final 6 using concatenation scheme to enhance orthogonality of basic (high-pH) and acidic (low-pH) separations and equalize concentration and distribution of peptides. Combined peptide fractions were evaporated under vacuum and resuspended at 250ng/μl in sample loading buffer (0.5% TFA, 2% ACN, 0.001% Zwittergent). Spike of iRT peptides was added to each of concatenated samples. Subsequently, 1μg of peptides from each fraction was analyzed by both the DDA and DIA methods and resulting data was used to generate appropriate library combination. Results of HpH fractionation are presented in Supplement 1 (Figure S5) . MS cycle time of 3180ms was matched across all the methods and chromatographic runs to give approx. 8 points per peak at FWHM (Supplement 1; Table S6) . For all DDA, STW and GPF-STW methods, polysiloxane at 445.12003 m/z was used as lock mass for real-time calibration. Ion-chromatogram libraries: To select optimal analysis approach, several versions of libraries were compiled. Each included combinations of appropriate fractions (GPF, HPH, collected in DDA and STW mode) and mixed sample runs (collected in STW mode). Spectronaut version 15.4.210913.50606-Rubin 40 , was used to search raw files with the build-in Pulsar engine against Homo sapiens Uniprot reference proteome FASTA file (UP000005640_9606, one sequence per protein). Pulsar standard settings were used, with following exceptions: digestion by Trypsin/LysC with P, carbamidometylation of C as fixed, acetylation of N-terminal, methylation of K or R, deamidation of Q or N, oxidation of M as variable modifications. Peptide length between 7 and 35 amino acids, max 5 variable modifications and max 2 missed cleavages per peptide. Spectra collected with staggered windows method were demultiplexed by build-in algorithm 39 . Dynamic mass calibration was used to correct both MS1 and MS2 spectra. Library was constrained to include from 3 to 6 of most intense b or y type fragments (300mz – 1800m/z range) per peptide of at least 3AA in length. Universal 1% FDR cutoff was used at the level of precursors; peptides and proteins. Retention times and RT window widths were calibrated using deep learning-assisted iRT regression included in Spectronaut. Results of library generation step are presented in Supplement 1 (Table S5) . Subsequently, mixed sample runs collected with standard 24m/z STW method (n=3) were searched with different versions of libraries to select best-performing combinations. The results are presented in Supplement 1 (Table S6) . 2.9 Data analysis: Biognosys Spectronaut was used to extract ion chromatograms from MS1 and MS2 data based on both the maximum ion peak intensity using dynamic mass tolerance, dynamic iRT-corrected retention time window and global, MS2-level signal normalization ( Supplement 1 Figure S3 ). Subsequently, data was searched with selected experiment-specific ion libraries. Q-value of 0.01 for both the precursors and peptides (corresponding to 1% FDR) was at the experiment level with a target-decoy approach using mutated sequences. For proteins Q-value as set at < 0.05 (5% FDR) and < 0.01 (1% FDR) at run-vise and experiment-vise levels, respectively. Quantification was performed at MS2 level, with Spectronaut bulid-in MaxLFQ algorithm. Within-sample peptide quantity was calculated as a mean of the peak areas of top 3 MS2 fragment ions, whereas protein quantity was expressed as a mean of top 3 peptides. For NGT and IGT group comparison, only the proteins present in at least 50% of the samples (Q-value percentile 0.5), with 2 unique peptides, presenting at least 50% expression difference (-0.585³ log 2 (fold change) ³ 0.585) and FDR-adjusted p-value <0.05 (Q-value 1.13) were regarded as significant 40 . For molecular pathway and molecular interaction analysis, IGT/NGT protein Log2 expression rations and Q-values were uploaded to Quiagen IPA 41 . Both expression Q-value cutoffs were used to restrict dataset to only significantly affected proteins (as mentioned above). IPA Core Analysis was restricted to mammalian proteins (human, rat, mouse) present in skeletal muscle, using stringent filtering for both molecules and interactions. Z-scores for pathways, diseases and bio functions were calculated based on expression log2 ratios. Corresponding -log 10 (p-value) of overlap were calculated using right-tailed Fisher's exact test with and FDR correction. Compare analysis function in IPA was used to correlate IPA results from IGT/NGT comparison generated with the HpH, GPF and DirectDIA TM approach. Results of comparison analysis at molecular pathway, disease and bio-function and individual protein level are presented in Supplement 2 (Table S3-S5) . Detailed results of IPA analysis performed on samples quantified with respective library are presented in Supplement 2 (Table S6-S8) . Protein functional clustering was performed with String 42 . We conducted a standard gene set analysis of significantly affected proteins, using their gene names and ranked FDR-adjusted p-values (4 ranks) expressed as -log 10 (p-value). Analysis was performed with full STRING Homo sapiens database (text mining, experiments, databases, co‑expression, neighborhood, gene fusion, co‑occurrence), confidence of supporting data was used as node connection (network edge) and minimum confidence of interaction was set at 0.500 (medium-high). Disconnected nodes were discarded. Clustering was performed using k-means and minimum number of clusters of 3. K-parameter (cluster number) was automatically addressed. Functional network enrichment was performed with FDR £ 1% and strength ³ 0.75. The ranked -log 10 (p-value) significance of change in IGT/NGT protein expression was used as node halo shading in network visualization. Full results of STRING cluster analysis are presented in Supplement 2 (Table S9-S11) . 3. Results 3.1 Utilization of m PAC column for LC/MS/HRMS proteomic analysis of skeletal muscle biopsy. The use of 50cm semiconductor-technology micro pillar array columns (mPAC) dictated separate gradient optimization steps, as the direct employment of gradients optimized for grain-based 50 microcapillary columns gave sub-par identification values. Best results were obtained with the use of multi-step (Table S2), 300nl/min 120min gradient (Table S3) at column load of 750ng of skeletal muscle digest (Table S4). Median CV% of retention time variation for all sample-spiked iRT peptides equaled to 26s, with FWHM of 17s (Figure S1 A and B). Pillar array column delivered good chromatographic resolution, retention time stability and reproducible detector signal (Figure S1C and D) accompanied by exceptionally low operating pressure as compared to typical, high-performance, particle-based capillary chromatography column of comparable length (Figure S1D). 3.2 Initial comparison of methods for ion-chromatogram library generation. At the early stage of ion-chromatogram library construction, we compared total number of identifications at peptide and protein level for mixed, study-wide muscle sample. HpH fractionation with fraction concatenation (Supplement 1, Figure S6) yielded highest identification numbers, at the level of both the individual and combined fractions. HpH outperformed in this regard both GPF (Supplement 1, Figure S5) and DirectDIA TM (Supplement 1, Figure S7) performed on 6 constitutive samples. As the results from HpH and GPF fractionation include runs from 6 independent fractions, we compared them with the values from 6 constitutive non-fractionated sample runs analyzed with DirectDIA TM approach. Moreover, to observe the impact of the number of analyzed samples on total number of identifications reported by DirectDIA TM we performed 12 constitutive injections, which yielded results comparable to GPF approach (Supplement 1, Figure S5 and S7A). 3.2 Different approaches to ion-chromatogram library construction for the proteomic analysis of skeletal muscle biopsy. Subsequently, we assembled a number of ion-chromatogram libraries employing only representative fractionation and analysis type techniques or in combination with additional non-fractionated study-wide mixed sample runs (Supplement 1, Table S5), performed with DDA or DIA method. The inclusion of those non-fractionated study-wide mixed runs yielded fractionated/non-fractionated hybrid libraries, with improved retention time alignment across the fractions. Addition of those samples significantly decreased the width of extracted ion-chromatogram (XIC) windows within libraries and improved overall identification number at PSM, peptide and protein level (Supplement 1, Table S5). At this step, we selected HpH-fractionated, DDA-acquired hybrid library, supplemented with 3 non-fractionated DDA samples (HpH/DDA-H library); GPF-fractionated, STW-acquired hybrid library, supplemented with 3 non-fractionated DDA-acquired samples (GPF/STW-H library) and DirectDIA TM approach for subsequent analysis (Figure 1, Supplement 1, Table S5). To further evaluate usefulness of selected approaches in the analysis of skeletal muscle proteome, we compared libraries at the level of PSMs, peptides and proteins. HpH/DDA-H hybrid library excelled in the number of unique PSMs (13771 out of 34740 total), which translated into highest number of unique peptides (8997 out of 23372 total) and proteins (952 out of 2515 total) identified within this library (Figure 2A to C, Supplement 1 Stable S5). Second-best GPF/STW-H library included 44 unique proteins (out of 1573 total), whereas DirectDIA TM only 39 (out of 1138 total). Both libraries were almost fully contained within HpH/DDA-H library at the level of peptides and proteins (Figure 2A to C). HpH/DDA-H library included higher number of low-intensity PSMs (Figure 2D), which possibly reflects peptide dilution and loss during fractionation procedure. GPF/STW-H library and DirectDIA TM approach yielded higher number of high-intensity PSMs, which in case of the latter could arise from prioritization of high-signal PSMs in de-novo identification pipeline and higher cut-off for low-intensity ones to decrease false-positives. At peptide physicochemical properties level, DirectDIA TM approach yielded slightly longer, more hydrophobic peptides, as reflected by their higher mean molecular weight, GRAVY score and hydrophobicity index as compared to both fractionation-based approaches. HpH/DDA-H library based on slightly shorter, hydrophilic peptides. This peptide characteristic could be result of chemical high-pH fractionation and retention of longer, hydrophobic peptides on ethylene bridged hybrid particles of C18 BEH column in high-pH conditions, absorption to plastic during sample transfers, and incomplete solubilization after vacuum concentration characteristic to HpH fractionation pipeline. GPF/STW-H yielded peptides with physicochemical characteristic in between those observed for DirecDIA TM and HpH/DDA-H regarding both the peptide length and hydrophobicity (Figure 2E). 3.4 Evaluation of the ion-chromatogram libraries and DirectDIA TM approach for the analysis of diabetes and exercise-related changes in skeletal muscle. To further evaluate usefulness of respective approaches for the analysis skeletal muscle proteome we performed pathway enrichment analysis using IPA software on proteins identified within a given library, with special emphasis on pathways connected to diabetes, muscle function and energy metabolism. IPA was able to identify 618 significantly enriched pathways in HpH/DDA-H library, covering 37% of their molecular members (pathway fraction) with mean value of 27 proteins per pathway (Figure 2F insert). While GPF/STW-H library presented similar values to HpH/DDA-H ones, DirectDIA TM approach yielded considerably lower values for the number of enriched pathways, mean enrichment -log 10 (p-value), pathway fraction and mean proteins per pathway (Figure 2F insert). Detailed enrichment analysis of insulin resistance, protein, energy metabolism, aging and muscular physiology-related pathways reflected above metrics, with HpH/DDA-H library containing highest number of proteins involved – among others - in mitochondrial metabolism, MAPK/AMPK and PI3K/AKT signaling, protein ubiquitination, autophagy, EIF2/eIF4 -controlled protein synthesis and NRF2-mediated oxidative stress response (Figure 2F). GPF/STW-H library yielded second-best results for mitochondrial metabolism, eIF4/p70S6K signaling and protein synthesis/degradation-related pathways, whereas DirecDIA TM approach presented lowest number of proteins and enrichment significance in all of the analyzed pathways (Figure 2F). We concluded library description stage of our study with the replicate (n=3) analysis of mixed, study-wide skeletal muscle sample to observe the impact of respective libraries on PSM, peptide and protein identification metrics. Compared to other non-hybrid library combinations, inclusion of mixed samples decreased by approx. 50% chromatographic XIC window width used for localization of MS/MS peptide fragments (from 5.1 min. to 2.9 min in case of GPF/STW-H library), which translated into higher identification scores noted for all hybrid libraries (Supplement 1, Table S6). Mean number of peptides per protein identified in mixed muscle samples ranged from 8.1 (DirectDIA TM ) to 9 (HpH/DDA-H). Approx. 94% of proteins present in GPF/STW-H library were identified in mixed muscle samples (Library recovery, Table S6), and each mixed sample presented 92% of total number of proteins identified in triplicate (Completeness, Table S6). Peptide MS/MS spectra from GPF/STW-H library were able to explain approx. 61% of total ion chromatogram signal (Explained TIC, Table S6). For HpH/DDA-H approach, the values for library recovery, sample completeness and explained TIC equaled to 62%, 90% and 65%, respectively, whereas DirectDIA TM presented 100%, 100% and 61% for the above parameters. I this case, highest performance at sample completeness and library recovery can be attributed to the nature of this approach, which creates internal ion-chromatogram library from all the samples included in the experiment. This ensures the recovery of complete set of IDs from each sample, albeit with overall lower number of IDs compared to deeper, fractionation-based libraries. Surprisingly, although HpH/DDA-H outperformed other approaches as for muscular proteome coverage at library assembly stage, analysis of skeletal muscle samples with GPF/STW-H library yielded highest unique PSM, peptide and protein identifications, as compared to both the HpH/DDA-H library and DirectDIA TM approach (Figure 3A to C, Table S6). Physicochemical characteristics of peptides identified within skeletal muscle samples was similar to the one observed for whole libraries, with HpH and DirectDIA TM displaying identification bias towards shorter, hydrophilic and longer, hydrophobic peptides, respectively (Figure 3D). To estimate reproducibility of respective approaches, we analyzed CV% distribution at PSM, peptide and protein level. Surprisingly, most direct DirectDIA TM approach without fractionation and library building was characterized by highest CV% at each level of the assay, as compared to both HpH/DDA-H and GPF/STW-H (Figure 3E). At final, protein level, median CV% equaled 15.1% for DirectDIA TM , 11% for GPF/STW-H, and 11% for HpH/DDA-H. Ultimately, GPF/STW-H approach yielded highest precision of the assay, with 70% of all proteins measured with CV% below 20%, compared with 69% for HpH/DDA-H and 69 for DirectDIA TM (Figure 3E). Protein rank distribution analysis of respective approaches revealed that GPF/STW-H library was able to identify and quantify higher number of proteins at all abundance levels in skeletal muscle samples as compared to other approaches (Figure 3F), with HpH/DDA-H library yielding similar, yet inferior results. Analysis performed with fractionation-based libraries gave significantly (numerically) better results regarding protein rank distribution, which was especially visible for low-abundance proteins (Figure 3F). Quantitative measurements performed with respective approach at displayed good, significant reciprocal correlation (Pearsons r>0.9 with p<0.00001 in all cases) at both the PSM, peptide and protein level. At final, protein level best correlation was observed for GPF/STW-H vs DirectDIA TM data (Pearsons r=0.9331, p<0.00001), which could reflect their more direct nature of measurement, compared to chemical fractionation-based HpH/DDA-H approach of library construction. To estimate the feasibility of the particular library in the detection of the muscular proteome alterations induced by glucose intolerance, we performed pathway enrichment analysis using proteins identified by a particular approach. It revealed, that samples analyzed with GPF/STW-H library yielded highest number of significantly enriched pathways (Supplement 1, Figure S8 - table insert), although pathways identified by second-best HpH/DDA-H approach displayed slightly better results regarding the mean significance of enrichment (7.8 vs 7.7 -log 10 (p-value)) and mean number of members per enriched pathway (21.4 vs 20.1). Analysis of enrichment of insulin resistance-relevant pathways revealed, that pathways from both the fractionation-based approaches had similar enrichment score (as measured by -log 10 (p-value)) and were equally populated by their molecular members (as measured by the number of identified proteins; Supplement 1, Figure S8). Pathways involved in autophagy, NRF2-mediated oxidative stress response and calcium signaling presented highest enrichment and protein count scores in HpH/DDA-H approach, whereas EIF2-mediated translation control and sirtuin signaling were most pronounced in GPF/STW-H ( Supplement 1, Figure S8). 3.5 Application of fractionation-based library approach or DirectDIA TM in the analysis of skeletal muscle proteomic alterations in post-training skeletal muscle from prediabetic subjects. 3.5.1 Protein identification and quantitation metrics. Finally, we utilized each of the selected approaches to identify proteomic alterations invoked by glucose intolerance and insulin resistance in skeletal muscle samples from subjects which underwent structured mixed-mode exercise regimen. When applied to the whole experimental IGT/NGT sample set, both HpH/DDA-H and GPF/STW-H approaches were able to identify identical number of protein ratios (Supplement 1, Table S7, Figure S9A to C). Although DirectDIA TM excelled at the number of PSMs, it did not translate into higher number of peptides and proteins than those observed for both fractionation-based approaches HpH/DDA-H and GPF/STW-H (1218 vs 1456, respectively). DirectDIA TM displayed highest sample completeness (90% vs 64% and 84% for HpH/DDA-H and GPF/STW-H, respectively), library recovery (100% vs 86% and 99%, respectively), peptides per protein (9.8 vs 9.2 and 9.4, respectively), percentage of explained TIC (71% vs 68% and 57%, respectively) and narrowest XIC windows (3.2 min. vs 4.2 and 3.8, respectively). In most of the above quality-related determinants, GPF/STW-H presented second-best results compared to DirectDIA TM (Supplement 1 Table S7), simultaneously displaying similar or better numbers at PSM, peptide and protein fold change ratios than the HpH/DDA-H approach (Supplement 1 Table S7, Figure S9A to C). Comparison of identification results for proteins passing 2 unique peptides cutoff, p-value<0.05 cutoff (equivalent to FDR-corrected Q-value of 0.05) and 50% fold change requirement (equivalent to 0.585£-log 2 (FC)£-0.585), revealed that although GPF/STW-H was able to quantify greatest total number of IGT/NGT protein ratios at 2-peptide and p-value cutoff, HpH/DDA-H approach excelled in the number of unique identifications at each of the subsequent steps. Finally, HpH/DDA-H analysis yielded a total of 140 significantly affected proteins passing all of the cutoffs, compared to 131 for GPF/STW-H and 64 for DirectDIA TM (Supplement 1, Figure S9D to F). Protein abundance rank analysis also confirmed higher proficiency of both fractionation-based approaches in protein identification compared to DirectDIA TM (Supplement 1 Figure S9G), although the latter one was able to quantify 29 more IGT/NGT ratios at the lowest protein abundance compared to both the HpH/DDA-H and STW/DIA-H.Study-wide correlation analysis of protein abundance between respective approaches yielded high Pearson r values for both the IGT-only and NGT-only samples (r>0.92, p<0.00001 in all cases), with HpH/DDA-H and STW/DIA-H displaying highest correlation of respective protein expression (Supplement 1, Figure S10A). Correlation analysis of IGT/NGT differential protein expression (expressed as protein log2 fold change, log2FC) yielded significant (p0.55, all cases) and Q-value (Pearson r>0.7, all cases) cutoffs (Supplement 1, Figure S10B) . Introduction of 50% fold-change cutoff drastically reduced the number of proteins shared between appropriate approaches, yet yielded highest Pearson r values of 0.91 and 0.96 for STW/DIA-H vs DirectDIA TM and HpH/DDA-H vs DirectDIA TM , respectively (p<0.00001). Despite highly correlated results of protein expression between HpH/DDA-H and STW/DIA-H at the level of IGT-only and NGT-only samples, the final log2FC values for significantly affected proteins displayed modest, although significant Pearson correlation of 0.7 (Supplement 1, Figure S10B) . Importantly, when compared between approaches, the shared proteins passing all 3 cutoffs and those passing 2-peptide and p-value cutoff (with minor exceptions) displayed the same direction and similar degree of differential regulation (Supplement 1, Figure S10B) . 3.5.2 Molecular pathways enrichment analysis Subsequently, using significant-only proteins (passing all 3 cutoffs), we performed pathway enrichment analysis and protein functional clustering, to identify proteomic alterations evoked by prediabetic state in post-training muscle and to evaluate the usefulness of each approach in the identification of the above changes. Quantitatively, GPF/STW-H identified highest number of significantly affected pathways (233) compared to both HpH/DDA-H and DirectDIA TM (200 and 98, respectively). DirectDIA TM performed better than both fractionation-based approaches, regarding mean pathway enrichment and proteins/pathway metrics (Figure 4, table insert). All of the approaches were equally effective in the detection of alterations in EIF2-mediated protein synthesis in IGT group, yet only DirectDIA TM was able to identify significant changes in protein ubiquitination, ubiquitin-like FAT10 signaling and detected changes in autophagy pathway with higher sensitivity than other approaches (Figure 4A, Supplement 2 Table 6 to 8) . Higher sensitivity of DirectDIA TM was also noted for all of the studied Aging/Stress/ROS-related pathways, such as sirtuin pathway, BAG2-mediated stress response and NRF2 oxidative stress response. Compared to both fractionation-based techniques, DirectDIA TM was unable to identify changes in pathways related to skeletal muscle physiology and contractile function, in which GPF/STW-H approach displayed best performance (Figure 4A) . Interestingly, all techniques were unable to detect in IGT group alterations in molecular pathways commonly connected with insulin resistance and prediabetes, such as mitochondrial dysfunction, PKA and AMPK and Type II DM-related signaling, which suggest that above hallmarks of diabetic state were normalized in the muscle of trained prediabetics. Significant changes were observed in protein synthesis-related eIF4 and p70S6K pathways, glucocorticoid signaling, and ERK/MAPK signaling, yet only GPF/DIA-H approach was able to detect alterations in PI3K/AKT-controlled pathway. To detect directionality of the changes, we performed z-score analysis on the significantly affected proteins, which revealed that IGT post-training muscle displayed down-regulation (z-score<2) in a total of 7 molecular pathways responsible for control of the protein synthesis, compared to post-training NGT counterparts (selenoaminoacid metabolism, detection of amino-acid deficiency, translation initiation, elongation, termination, RNA processing and EIF2 signaling) (Figure 4B, Supplement 2 Table 6 to 8) . Both GPF/STW-H and DirectDIA TM were equally sensitive in the detection of the above alterations (with EIF2 pathway as an exception), whereas HPH/DDA-H approach identified only 3 out of 7 in total. Contrary to DirectDIA TM , both fractionation-based techniques detected significant down-regulation in the expression of proteins involved in skeletal muscle contraction (at the level of actin cytoskeleton, integrin signaling and muscle hypertrophy response, Figure 4B), whereas only DirectDIA TM was able to detect detrimental alterations in pathways responsible for DNA synthesis, replication and repair and PTEN-dependent regulation of the cell cycle (Figure 4B, Supplement 2 Table 6 to 8) . Only GPF/STW-H approach was sensitive enough to detect down-regulation in the PI3K/AKT, a key insulin signaling pathway. Regarding structural and functional alterations which distinguish post-training IGT group from their normoglycemic counterparts, both HpH/DDA-H and GPF/DIA-H methods detected greater number of changes compared to DirectDIA TM (Figure 4C, Supplement 2 Table 6 to 8) . Interestingly, whereas HpH/DDA-H was able to detect significant changes in lipid metabolism-related muscle functions (lean body mass retention, quantity of lipid droplets) and myofiber differentiation, GPF/DIA-H excelled in detection of all muscle contractile and morphology-related changes, such as altered Ca 2+ storage, abnormal morphology of muscle and its fibers, strength of contraction and disrupted motor plate function (Figure 4C, Supplement 2 Table 6 to 8) . DirectDIA TM detected molecular alterations connected with density of neuromuscular junctions and skeletal muscle fibrosis. 3.5.2 Protein functional clustering To further investigate the differences between IGT and NGT subjects observed in post-training skeletal muscle, and to identify relationships between differentially regulated proteins, we performed functional protein clustering analysis with each of the different approaches. Total of 103 out of 141 significantly affected proteins identified by HpH/DDA-H method (74% of total) generated 4 major interconnected clusters, all implicated in different aspects of protein metabolism (Figure 5, Supplement 2, Table S9) . Cluster I encompassed proteins involved in the regulation of DNA expression (MAPK kinase group), mRNA processing (HNRNP proteins), translation control (EIF factors of initiation, elongation and termination) and ribosomal subunits. Cluster II aggregated proteins involved in ER-mediated protein processing and post-translational modifications, whereas Cluster III and VI gathered proteins involved in Golgi-mediated vesicular transport and endocytic vesicular recycling. Those findings indicate that HPH/DDA-H was able to detect significant alterations in molecular control of protein synthesis, processing and trafficking in trained pre-diabetics compared to respective normoglycemic counterparts. This observation explains the presence of Cluster IV, composed of muscle contractile apparatus and extracellular matrix proteins, which suggest abnormal skeletal muscle composition in post-training IGT group compared to NGT one. Finally, HPH/DDA-H approach identified a known hallmark of prediabetes i.e. alterations in proteins connected with mitochondrial fatty acids metabolism (Cluster V). Similarly, clustering of 85 out of 130 (65%) of significantly affected proteins revealed by GPF/DIA-H emphasized alterations in mRNA processing, translation and protein metabolism (Cluster II and its 2 subclusters, which included MAPK kinases, EIF translation control factors, ribosomal subunits, RAS pathway members and aminopeptidases) and muscle contraction and extracellular matrix (Cluster III, which included light and heavy myosin chain isoforms, actin and EC matrix laminin B1, decorin and prolagin) (Figure 6, Supplement 2, Table S10) . Interestingly, unique finding of GFP/DIA-H approach was the identification of clusters, which molecular members are involved in branched amino-acids biosynthesis and degradation (Cluster IV e.g. mitochondrial BCAT2 branched chain aminotransferase, ACADBS short/branched chain acyl-CoA dehydrogenase, MCEE mitochondrial methylmalonyl-CoA epimerase), skeletal muscle carnosine metabolism (Cluster IV, e.g. ) and pyridoxal phosphate (vitamin B6) metabolism (Cluster V, i.e. PDXK pyridoxal kinase and PLPBP pyridoxal phosphate binding protein). Alterations in BCAA metabolism, muscular carnosine dipeptide content and vitamin B6 deficiency display strong correlation with prediabetes and subsequent progression toward Type 2 DM. Clustering of the most diverse protein set was noted for Cluster I (Figure 6, Supplement 2, Table S10) , and included proteins connected with gene expression and mRNA processing at mitochondrial and nuclear level (e.g. TFAM mitochondrial transcription factor A, proteins from HRNR heterogeneous nuclear ribonucleoprotein family, SNRBP RNA spliceosome protein and STRAP ribonucleoprotein assembly protein), cytoskeleton/plasma membrane interaction and assembly proteins (e.g. caveolae associated CAV1, CAVIN4, STIM1 proteins and cytoskeleton assembly and plasma membrane anchoring FLNA flaminin, MSN moesin, SNTB1 b-1-syntropin and ANK3 Ankyrin-3 proteins) lipid transport and metabolism-associated proteins (APOE, APOC1 lipoproteins and FABP4 fatty acids binding protein) and finally ROS metabolism and pentose phosphate pathway proteins (GPX3 glutathione peroxidase 3, TXN2 mitochondrial thioredoxin and H6PD hexose-6-phosphate dehydrogenase/glucose 1-dehydrogenase, PGLS 6-phosphogluconolactonase, respectively). APOE played the central, linking role in formation of Cluster I, possibly due to its impact on caveolae function, lipid metabolism, and mitochondrial dysfunction function (Figure 6, Supplement 2, Table S10). Protein metabolism-related clustering of significantly affected proteins was also noted for DirectDIA TM analyzed IGT vs NGT dataset. A total of 38 out of 64 proteins (65%) yielded 6 independent (not connected) clusters, with Cluster I, II and IV aggregating proteins involved in muscle contraction, cytoskeleton formation and extracellular matrix (Cluster I, e.g mysin and actin isoforms, collagens, serpin, laminin) mRNA processing, translation and protein degradation (Cluster II e.g) and protein folding and glycosylation (Cluster V e.g. calumenin, DDOST and RPN2 glycosyltransferases). Interestingly, Cluster II included both the molecular members of protein synthesis pathways (EIF factors, ribosomal subunits) and proteasomal degradation pathway (PSM proteasomal proteins) ( Figure 7, Supplement 2, Table S11) , which was unique finding of DirectDIA TM analysis. Cluster III connected proteins involved in muscular calcium binding and Ca 2+ regulation of muscle contraction (S100 family proteins, parvalbumin) whereas 2 members of Cluster V shared the same molecular function as acyl-CoA – synthesizing enzymes (mitochondrial medium and short-chain Acyl-CoA ligases). 4. Discussion Introduction of high-throughput proteomic analysis in clinical applications carries a promise of possible ground-breaking discoveries at the level of both the foundational science and medical applications. Initial attempts based on 2DGE with MALDI-TOF peptide identification or FT-ICR LC/MS ultra-high resolution mass spectrometry, although encouraging from the basic science standpoint, were lacking in important qualities, such as sample throughput, reproducibility and proteome coverage. Additionally, plasma – a readily-available clinical sample - presents significant analytical challenge regarding protein matrix composition (high-abundance protein “iceberg” effect). Those shortcomings were the result of several bottlenecks and variabilities at almost every step of the proteomic pipeline, including cumbersome sample preparation, consistency of in-house prepared nanoLC columns, slow-cycle DDA-based acquisition and early software employed in peptide identification and protein quantitation. Current generation of MS-based proteomic pipelines finally resolve all the above issues. Moreover, partial application of modern pipelines on previous-generation mass spectrometers allows for the high-throughput proteomic analysis of quality and reproducibility required for clinical applications. Taking in to account all of the above, we employed modern DIA-based approach consisting of staggered-window MS/MS acquisition, customized ion-chromatogram libraries and highly reproducible semiconductor-technology based micro pillar array columns coupled with Orbitrap-HRMS for the analysis of challenging skeletal muscle samples. Fine-tuning of mPAC chromatography, DIA acquisition and library construction allowed for elucidation of skeletal muscle proteome changes in prediabetic patients, confirming presence of persistent, adverse changes in muscle proteome, despite 3 months of structured exercise. Compared with basic science proteomics research, which can afford for multiple re-analysis attempts and longer analysis times to achieve best possible data quality, proteomics in clinical applications puts emphasis on robustness, batch-to-batch reproducibility and throughput of the complete analysis pipeline. To improve robustness and reproducibility of LC separations we employed micro pillar array columns which – compared to particle based counterparts – display decreased batch-to batch variability, significantly lower backpressures, full flow reversibility and low carryover 43 . Those features arise from semiconductor-type manufacturing process, which creates 5mm ID pillars, 18μm in height separated by 2.5μm gaps 27 . In our study, 1 st Gen 50cm mPAC column paired with matching 1cm mPAC trap generated 20x lower backpressure than particle-based counterpart (25bar vs 500bar) and negligible pressure drop during trap valve operation, increasing robustness of HPLC analysis. The column displayed good retention time stability, and was resistant to clogging even in case of samples rich in fibrous biological polymers (skeletal muscle). In case of increased backpressure, mPAC column could be reverse-flushed to restore its original performance, increasing column longevity. This cannot be applied to packed, particle-based nanoflow capillary columns without the risk of stationary phase loss. Currently, pillar array columns are increasingly used in proteomics applications 28,44-47 with 2 nd generation columns (2.5mm ID pillars, 25mm height 1.5μm interpillar distance) displaying improved characteristic regarding theoretical plate heigh and resolution 27,28 . Clinical applications can benefit from newest, rectangle pillar-based design generation (75mm x 3mm pillars), which despite its short 5.5cm length allows for higher-throughput, mLC-like analysis, with separation power comparable to its 50cm circular pillar array counterpart 44,48 . Low column-to-column variability and retention time consistency displayed by micro pillar array column are crucial in the robust implementation of DIA analysis in both the library-based and pure bioinformatic approaches. For our study we selected staggered windows approach (STW-DIA) for both the gas-phase fractionation library generation and experimental runs, first introduced by Searle, Pino and Amodei 37-39 . Although non-overlapped window placement would yield increased MS/MS cycle times, in our view the window staggering is better suited for older-generation quadrupole-orbitrap instruments, due to the correction of non-rectangular isolation characteristic of quadrupoles and increased precision of MS/MS measurements crucial for DIA-based assays. Despite being rarely employed, this approach is gaining increasing acceptance and was employed in recent studies on both previous generation and modern mass spectrometers 30,49 . To elucidate the impact of ion-chromatogram library assemble on the results of DIA-STW based analysis of exercise-induced changes in prediabetic subjects, we selected 3 approaches with increasing deepness of the library coverage, namely High-pH fractionation (HpH), gas-phase fractionation (GPF) and pure computational approach (Spectronaut TM DirectDIA TM ). Post-digestion high-pH fractionation with fraction concatenation (HpH) was shown multiple times to generate most comprehensive ion-chromatogram libraries, as compared to other protein and peptide fractionation techniques, such as PAGE, IEF focusing variants, HILIC, SCX and SAX ion exchange 50-52 . When performed at microflow ranges (50-100μl/min) with 100μg of total protein digest, we were able to perform all the necessary steps such as fraction, fraction concatenation, vacuum concentration and subsequent LC/MS/MS analysis in a single 96 well-plate format, which assures low sample loss. Nevertheless, HPLC-based physical fractionation is time-consuming, requires additional instrumentation (vacuum concentrators, dedicated HPLC with fraction collection or multimode nanoLC/microLC HPLC system etc.). In-source gas-phase fractionation offers cost and time-effective alternative to physical peptide fractionation, requiring only several additional analysis runs, albeit with lower library coverage. To further increase the deepness proteome analysis, staggered-windows gas-phase fractionation was recently combined by Penny et al. with ion-mobility technique (TIMS-diaPASEF), yielding significant improvements over its basic versions 53 . Finally, pure bioinformatic solutions to DIA analysis rely on direct identification of peptide product ions without the prior assembly of ion-chromatogram libraries. Currently, this approach to DIA analysis displays fastest growth in capabilities and multitude of available software solutions, presenting different analytical approaches (e.g spectrum-centric, peptide-centric, in-silico fragmentation libraries etc.). Recently, pure computational DIA analysis was updated with machine learning, neural networks and AI capabilities (for excellent reviews and comparative studies see 30,54 and 33,55 , respectively). Direct analysis of DIA-acquired MS/MS runs presents the fastest, direct approach, yet requires significant computational resources. Moreover, contrary to both the fractionation-based libraries which, ideally, should be prepared with the use of study-wide mixed sample, it can be applied continuously, in parallel to sample collection and processing before the completion of the study. Our primary objective was to identify the differences of 3 major approaches on the outcome of DIA-based analysis of muscular proteome changes evoked by exercise in prediabetic patients. We hypothesized, that assembly of spectral libraries can have direct impact on the outcome of the analysis and the final biological findings. Firstly, we selected best performing libraries among several possible combinations. Interestingly, hybrid libraries which in addition to GPF or HpH fraction runs included full mass range DDA runs from study-wide mixed samples displayed narrower XIC search windows, which translated into lower CV% of replicate analysis. Greatest improvement was noted for both of the fractionated libraries, which in our opinion is connected with the improved RT alignment of across particular library fractions, as only non-fractionated samples contained all of the proteins and peptides. Surprisingly, inclusion of non-fractionated study-wide sample had greater impact on RT alignment and XIC window width, than the presence of iRT peptides in each of the fractions, which in theory should provide RT anchor points across library runs. Although this improvement should be most visible in the case of GPF hybrid library (as iRT peptides fall in different mass fractions), it was also noted for HpH hybrid library, where iRT peptides were added to concatenated fractions prior to LC/MS runs. As expected, DirectDIA TM approach displayed the best RT alignment and narrowest XIC windows, yet yielded greatest CV% of replicate analysis. This outcome was somehow surprising, as protein rank distribution and precursor intensity distribution suggested, that DirectDIA TM displayed bias toward higher intensity signals, inherently easier to quantify. Moreover, DirectDIA TM analysis targeted longer, more hydrophobic peptides, compared to analysis performed with HPH/DDA-H library, which could be explained by hydrophobic peptide loss during HpH fractionation. Yet the difference was also visible between DirectDIA TM and GPF/STW-H analysis, which cannot be explained by mass-range gas-phase fractionation. We observed this phenomenon in both the results of the analysis of mixed-sample replicates and whole sample. Those findings suggest, that biological outcomes of the analysis performed with different approaches could differ due to targeting of peptides with different physicochemical properties, displaying slight bias toward membrane or soluble proteins. As expected, HpH/DDA-H ion-chromatogram library contained the highest number of identified proteins and presented most populated pathways relevant to T2D and insulin resistance. Yet this advantage did not translate into higher-quality results, when employed to quantify proteins in study-wide mixed sample. The GPF/STW-H – based results were similar or better regarding quantity of proteins and their number included in T2D-relevand pathways. This gives significant advantage for staggered-windows gas-phase fractionation approach, which requires less time and resources for implementation. Regarding the outcome of the analysis performed on whole set of experimental samples, we noted both similarities and significant differences depending on the use of particular approach. Although we observed strong, group-wise correlations between individual proteins in IGT or NGT groups quantified with 3 different approaches (Figure S10A), the final IGT vs NGT differential expression ratios displayed weaker association (Figure S10B). Moreover, each approach turned up relatively different set of proteins, when all significance cut-offs were considered (>= 2 peptides, Q-value<0.05, 50% FC), with only 8 common proteins (down-regulated in IGT group: TAGLN, MYBPH, KRT2, S100A13, PRELP ACADSB and up-regulated in IGT group C4A and ACTN3) common between all 3 approaches. Those discrepancies arise from differences in FDR-corrected p-values (Q-values) and calculated differential expression ratios, as those parameters substantially decreased the number of shared proteins. Interestingly, both the molecular function and the direction of regulation of some of those proteins align with the metabolic and functional deficiencies observed in insulin-resistant muscle. Down-regulation of MYBPH, which is highly expressed in insulin-sensitive 56,57 , mitochondria-rich type 1 oxidative muscle fibers 58 , can reflect the decrease of fiber type in IGT group muscle, despite structured exercise regimen. Similarly, down-regulation of S100A13 calcium-binding protein, which over-expression is connected with mitochondrial biogenesis in hypoxia-trained skeletal muscle 59 , aligns with decreased oxidative capacity and mitochondrial content observed in insulin resistant muscle 60 , whereas decrease in expression of mitochondrial ACADSB branched-chain dehydrogenase can be traced to both mitochondrial deficiency 61 and disrupted branched chain AA metabolism in prediabetic subjects 62,63 . Regarding over-expressed proteins, complement protein C4A up-regulation corresponds with increased pro-inflammatory response observed in obesity-induced insulin resistance 64,65 . Taking into account all significantly affected proteins, each approach was more sensitive towards particular changes observed in skeletal muscle proteome of IGT group. DirectDIA TM analysis identified pathways involved in proteasomal protein degradation, cell cycle control and DNA replication and repair, which was not observed in both fractionation-based approaches. Moreover, DirectDIA TM was more sensitive towards identification of changes connected with protein synthesis and degradation, oxidative stress and sirtuin signaling pathways, although all approaches signaled decreased expression of proteins involved in translation control. GPF/STW-H library based analysis was more sensitive towards detecting changes in skeletal muscle calcium signaling and morphology, whereas HpH/DDA-H in eIF4 and p70S6 kinase signaling and decreased muscle hypertrophy. Differences in biological data interpretation were also visible at the level of protein functional clustering, with DirectDIA TM detecting unique alternations in proteins involved in PTM glycation, Ca 2+ binding and acyl-CoA synthesis, GPF/STW-H in proteins involved in branched amino-acid metabolism and vitamin B6 metabolism, whereas HpH/DDA-H in ER protein processing and Golgi vesicular transport. All of the approach-dependent unique protein clusters are important for metabolic function of skeletal muscle in normoglycemia, as alternations in muscular Ca 2+ metabolism 66,67 and protein glycation 68 , branched-chain AA 62,63 and vitamin B6 metabolism 69,70 and proteins-synthesis related ER stress 71,72 are hallmarks of insulin resistance and T2DM. Nevertheless, significant alternation observed in trained IGT group as compared to respective NGT group, was the down-regulation of pathways involved in protein synthesis, despite 3 months of controlled exercise regimen. Adverse effects of insulin resistance on skeletal muscle protein synthesis could be responsible for the observed decrease in sarcomere and extracellular matrix proteins, leading to skeletal muscle contractile dysfunction 17,18 . Regarding both the qualitative and quantitative aspects of particular approach, HpH/DDA-H – based analysis identified the greatest number of significant proteins and generated unique protein interaction networks, covering important aspects of skeletal muscle metabolism such as ER stress and Golgi protein processing. On the other hand GPF/STW-H approach presented similar quantitative performance, identified significant portion of affected proteins and unique features of muscular insulin resistance (disturbance in branched AA metabolism), being less problematic to implement. Finally, libraryless DirectDIA TM analysis, although presented fewer significantly affected proteins, was able to robustly detect common features of skeletal muscle insulin resistance, e.g. disturbed control of protein synthesis and proteasomal protein degradation. Taking all of the above, gas-phase fractionated library, acquired in staggered windows mode and supplemented with DDA full mass range runs constitute an attractive alternative for time- and resource-consuming physical peptide fractionation. Moreover, identification and quantitation of whole sample set using GPF/STW-H approach was computationally less demanding than pure computational approach, moreover doubling the number of significant proteins. Although improvements in pure computational identification and quantitation will narrow the gap between library-based and libraryless approaches, the similar augmentation of library-based quantitation will further advance DIA-based assay. 5. Conclusions Among multitude of MS acquisition modes employed in proteomic assays, data-independent analysis coupled with robust chromatographic separation is best suited for large projects performed on clinical samples. Important aspect of DIA-based assay is the assembly of ion-chromatogram libraries or selection of pure computational strategy for DIA data interrogation, which hypothetically can lead distinct biological interpretations. Our goal was to employ different strategies to library construction to observe their impact on the DIA-based analysis of post-exercise differences in skeletal muscle proteome between glucose intolerant and healthy subjects. Although all tested approaches were able to detect alternations in control of protein synthesis and sarcomere protein expression, each one identified unique changes important to metabolic and contractile muscle function. In our view, staggered-windows gas-phase fractionated ion-chromatogram library presented the best balance between reproducibility, detection of significant changes, depth of analysis and ease of implementation. Combined together, the biological findings indicate, that despite extensive structured exercise regimen, insulin-resistant muscle displays disturbed molecular pathways implicated in protein synthesis, intracellular trafficking and processing, branched amino-acids and acyl-CoA mitochondrial metabolism and calcium balance, which can be the cause of muscular contractile dysfunction observed in insulin resistant subjects. Abbreviations 2DPAGE – 2 dimensional polyacrylamide gel electrophoresis ABC - Ammonium Bicarbonate ACG - Automatic Gain Control ACN - Acetonitrile AMPK - AMP-Activated Protein Kinase AU - Arbitrary Units AUC - Area Under the Curve BCA - Bicinchoninic Acid BCAA - Branched-Chain Amino Acids BMI - Body Mass Index Ca²⁺ - Calcium Ion Chol - Cholesterol DDA - Data-Dependent Acquisition DIA - Data-Independent Acquisition DirectDIA™ - Direct (libraryless) Analysis of Spectra form Data-Independent Acquisition DPP4 - Dipeptidyl Peptidase 4 EIF2 - Eukaryotic Initiation Factor 2 ER - Endoplasmic Reticulum FA - Formic Acid FDR - False Discovery Rate FT-ICR LC/MS - Liquid chromatography/Fourier transform ion cyclotron resonance mass spectrometry FPG - Fasting Plasma Glucose GLP1 - Glucagon-Like Peptide 1 GO - Gene Ontology GPF - Gas-Phase Fractionation HbA1c - Hemoglobin A1c HDL - High-Density Lipoprotein Cholesterol HOMA2 %B - Homeostatic Model Assessment of Beta Cell Function HOMA2 %S - Homeostatic Model Assessment of Insulin Sensitivity HOMA2-IR - Homeostatic Model Assessment for Insulin Resistance HOMAD - Homeostatic Model Assessment for Diabetes HpH – Reverse-phase HPLC Peptide Fractionation in high-pH conditions HRMS - High-Resolution Mass Spectrometry IAA - Iodoacetamide IFG - Impaired Fasting Glucose IGT - Impaired Glucose Tolerance INS - Insulin iRT – Stable Isotope-labelled peptides for Peptide Indexed Retention Time calculation LC/MS/MS - Liquid Chromatography/Tandem Mass Spectrometry LDL - Low-Density Lipoprotein Cholesterol LN₂ - Liquid Nitrogen MALDI-TOF - Matrix-assisted laser desorption/ionization – time of flight mass spectrometry MSX - Multiplexed MS/MS NCE - Normalized Collision Energy NGT - Normal Glucose Tolerance OGTT - Oral Glucose Tolerance Test PSM - Peptide-Spectrum Match PTM/PTMs - Post-Translational Modification/Post-Translational Modifications SDC - Sodium Deoxycholate SGLT2 - Sodium-Glucose Co-Transporter 2 SMM - Skeletal Muscle Mass STRING - Search Tool for the Retrieval of Interacting Genes/Proteins STW - Staggered Windows T2DM - Type 2 Diabetes Mellitus TCEP - Tris(2-carboxyethyl)phosphine TFA - Trifluoroacetic Acid TG - Triglycerides TIC - Total Ion Chromatogram VAT_DXA - Visceral Adipose Tissue measured by Dual-Energy X-ray Absorptiometry VO₂max - Maximal Oxygen Uptake μPAC™ - Micro Pillar Array Column Declarations Author Contribution A.C., L.S. and P.Z designed the study; L.S supervised subjects-related parto of the study and sample collection stage; A.C-R and P.K assisted in sample anthropometric data collection; A.C. and P.Z performed all sample analysis, data analysis, data interpretation and statistical analysis; M.C. assisted in data analysis nad statistical analysis; A.C and P.Z. wrote original draft manuscript and prepared all manuscript figures; A.B-Z . supplied resources, performed all diabetes-related data interpretation; P.Z., L.S. and A.K. supplied funding. All authors reviewed the manuscript. Data availability: The mass spectrometry proteomics data have been deposited to the ProteomeXchange Consortium (http://proteomecentral.proteomexchange.org) via the PRIDE partner repository 73 with the dataset identifier PXD055536. Reviewer account details: Project accession: PXD055536 Username: [email protected] Password: pIpjyglKQ2WB This study was supported by funds from the Ministry of Education and Science of Poland, within the project “Excellence Initiative—Research University”, Ministry of Health of Poland within the project “Center for Artificial Medicine at the Medical University of Bialystok” and the Medical University of Bialystok grant B.SUB.24.393 References D, V. et al. Projecting the economic burden of type 1 and type 2 diabetes mellitus in Germany from 2010 until 2040. Popul. health metrics . 22 10.1186/s12963-024-00337-x (2024). ED, P. et al. Economic Costs of Diabetes in the U.S. in 2022. Diabetes care . 47 10.2337/dci23-0085 (2024). 2. Diagnosis and Classification of Diabetes: Standards of Care in Diabetes-2024. Diabetes care . 47 10.2337/dc24-S002 (2024). Dludla, P. V. et al. Pancreatic β-cell dysfunction in type 2 diabetes: Implications of inflammation and oxidative stress. World J. Diabetes . 14 10.4239/wjd.v14.i3.130 (2025). Mezza, T. et al. β-Cell Fate in Human Insulin Resistance and Type 2 Diabetes: A Perspective on Islet Plasticity. Diabetes 68 , 1121. 10.2337/db18-0856 (2025). A, D. et al. Pharmacotherapy of type 2 diabetes: An update and future directions. Metab. Clin. Exp. 137 10.1016/j.metabol.2022.155332 (2022). KM, P. et al. Clinical characteristics, complications, comorbidities and treatment patterns among patients with type 2 diabetes mellitus in a large integrated health system. BMJ open. diabetes Res. care . 3 10.1136/bmjdrc-2015-000093 (2015). KI, G. et al. Interventions for Reversing Prediabetes: A Systematic Review and Meta-Analysis. Am. J. Prev. Med. 62 10.1016/j.amepre.2021.10.020 (2022). Borges-Canha, M. et al. Prediabetes remission after bariatric surgery: a 4-years follow-up study. BMC Endocr. Disorders . 24 10.1186/s12902-024-01537-0 (2025). H, Z. et al. Exercise training modalities in prediabetes: a systematic review and network meta-analysis. Front. Endocrinol. 15 10.3389/fendo.2024.1308959 (2024). P, M., Z, S. H., F, A. & B. & Dietary pattern scores in relation to pre-diabetes regression to normal glycemia or progression to type 2 diabetes: a 9-year follow-up. BMC Endocr. disorders . 23 10.1186/s12902-023-01275-9 (2023). T, S. et al. Quantifying the Relationship Between Physical Activity Energy Expenditure and Incident Type 2 Diabetes: A Prospective Cohort Study of Device-Measured Activity in 90,096 Adults. Diabetes care . 46 10.2337/dc22-1467 (2023). W, Y. et al. Different levels of physical activity and risk of developing type 2 diabetes among adults with prediabetes: a population-based cohort study. Nutr. J. 23 10.1186/s12937-024-01013-4 (2024). JA, F., BR, H., JK, R. J. P., LS, N. & S. & Changes in Cognition and Brain Function After 26 Weeks of Progressive Resistance Training in Older Adults at Risk for Diabetes: A Pilot Randomized Controlled Trial. Can. J. diabetes . 47 10.1016/j.jcjd.2023.01.004 (2023). M, L., MV, M. & A, K. & The effects of exercise treatment on learning and memory ability, and cognitive performance in diet-induced prediabetes animals. Sci. Rep. 10 10.1038/s41598-020-72098-0 (2020). HH, L., C, Y. Z., ZJ, G. & C. & Meta-analysis of the effect of exercise intervention on cognitive function in elderly patients with type 2 diabetes mellitus. BMC Geriatr. 24 10.1186/s12877-024-05352-z (2024). JW, S., AR, H. & SK, H. Greater Lower Limb Fatigability in People with Prediabetes than Controls. Med. Sci. Sports. Exerc. 52 10.1249/MSS.0000000000002238 (2020). M, D. & A, K. Investigating the Effects of Diet-Induced Prediabetes on Skeletal Muscle Strength in Male Sprague Dawley Rats. Int. J. Mol. Sci. 25 10.3390/ijms25074076 (2024). EM, R. et al. Reduced skeletal muscle phosphocreatine concentration in type 2 diabetic patients: a quantitative image-based phosphorus-31 MR spectroscopy study. Am. J. Physiol. Endocrinol. Metab. 315 10.1152/ajpendo.00426.2017 (2018). E, F. et al. Insulin Resistance Is Associated With Reduced Mitochondrial Oxidative Capacity Measured by 31P-Magnetic Resonance Spectroscopy in Participants Without Diabetes From the Baltimore Longitudinal Study of Aging. Diabetes 66 10.2337/db16-0754 (2017). A, H. et al. Human skeletal muscle mitochondrial dynamics in relation to oxidative capacity and insulin sensitivity. Diabetologia 64 10.1007/s00125-020-05335-w (2021). L, S. et al. The Response of Mitochondrial Respiration and Quantity in Skeletal Muscle and Adipose Tissue to Exercise in Humans with Prediabetes. Cells 10, (2021). 10.3390/cells10113013 K, O. Skeletal muscle proteomics: current approaches, technical challenges and emerging techniques. Skelet. Muscle . 1 10.1186/2044-5040-1-6 (2011). DT, C., AS, D. & R, M.-J., JP, Q. & Mass spectrometry-based proteomics approaches to interrogate skeletal muscle adaptations to exercise. Scand. J. Med. Sci. Sports . 34 10.1111/sms.14334 (2024). Gonzalez-Freire, M. et al. The Human Skeletal Muscle Proteome Project: a reappraisal of the current literature. J. Cachexia Sarcopenia Muscle . 8 (5). 10.1002/jcsm.12121 (2025). SJ, H., BN, S., CA, S. & JG, B. The application of proteomics in muscle exercise physiology. Expert Rev. Proteomics . 17 10.1080/14789450.2020.1879647 (2020). B, V. & G, D. J, O. d. B. On-Chip Comparison of the Performance of First- and Second-Generation Micropillar Array Columns. Analytical chemistry 95, (2023). 10.1021/acs.analchem.3c01829 Stejskal, K. et al. Deep Proteome Profiling with Reduced Carryover Using Superficially Porous Microfabricated nanoLC Columns. Anal. Chem. 94 , 15930. 10.1021/acs.analchem.2c01196 (2025). Barkovits, K. et al. Reproducibility, Specificity and Accuracy of Relative Quantification Using Spectral Library-based Data-independent Acquisition. Mol. Cell. Proteomics: MCP . 19 , 181. 10.1074/mcp.RA119.001714 (2025). K, F. et al. Data-Independent Acquisition: A Milestone and Prospect in Clinical Mass Spectrometry-Based Proteomics. Mol. Cell. proteomics: MCP . 23 10.1016/j.mcpro.2024.100800 (2024). BC, S. et al. Generating high quality libraries for DIA MS with empirically corrected peptide predictions. Nat. Commun. 11 10.1038/s41467-020-15346-1 (2020). Govaert, E. et al. Comparison of fractionation proteomics for local SWATH library building. Proteomics 17 , 1700052. 10.1002/pmic.201700052 (2025). R, L. et al. Benchmarking commonly used software suites and analysis workflows for DIA proteomics and phosphoproteomics. Nat. Commun. 14 10.1038/s41467-022-35740-1 (2023). L, S. et al. Metabolomic Profile of Skeletal Muscle and Its Change Under a Mixed-Mode Exercise Intervention in Progressively Dysglycemic Subjects. Front. Endocrinol. 12 10.3389/fendo.2021.778442 (2021). RA, S. et al. Human skeletal muscle biopsy procedures using the modified Bergström technique. J. visualized experiments: JoVE . 10.3791/51812 (2014). IR, L., ON, V. S., RR, S. & J. & Quantitative assessment of in-solution digestion efficiency identifies optimal protocols for unbiased protein analysis. Mol. Cell. proteomics: MCP . 12 10.1074/mcp.M112.025585 (2013). BC, S. et al. Chromatogram libraries improve peptide detection and quantification by data independent acquisition mass spectrometry. Nat. Commun. 9 10.1038/s41467-018-07454-w (2018). LK, P., SC, J., MJ, M. & BC, S. Acquiring and Analyzing Data Independent Acquisition Proteomics Experiments without Spectrum Libraries. Mol. Cell. proteomics: MCP . 19 10.1074/mcp.P119.001913 (2020). D, A. et al. Improving Precursor Selectivity in Data-Independent Acquisition Using Overlapping Windows. J. Am. Soc. Mass. Spectrom. 30 10.1007/s13361-018-2122-8 (2019). R, B. et al. Optimization of Experimental Parameters in Data-Independent Mass Spectrometry Significantly Increases Depth and Reproducibility of Results. Mol. Cell. proteomics: MCP . 16 10.1074/mcp.RA117.000314 (2017). A, K., S, T. & J, G., J, P. & Causal analysis approaches in Ingenuity Pathway Analysis. Bioinf. (Oxford England) . 30 10.1093/bioinformatics/btt703 (2014). D, S. et al. The STRING database in 2023: protein-protein association networks and functional enrichment analyses for any sequenced genome of interest. Nucleic Acids Res. 51 10.1093/nar/gkac1000 (2023). Rozing, G. Micropillar array columns for advancing nanoflow HPLC. Microchem. J. 170. http://dx.doi.org/10.1016/j.microc.2021.106629 (2021). Matzinger, M. et al. Micropillar arrays, wide window acquisition and AI-based data analysis improve comprehensiveness in multiple proteomic applications. Nat. Commun. 15 , 1019. 10.1038/s41467-024-45391-z (2025). K, S., K, M. & J, O. d. B., G, D., P, J. & Ultrasensitive NanoLC-MS of Subnanogram Protein Samples Using Second Generation Micropillar Array LC Technology with Orbitrap Exploris 480 and FAIMS PRO. Anal. Chem. 93 10.1021/acs.analchem.1c00990 (2021). BG, C., Y, M. & P, J., M, G., S, G. & Using micro pillar array columns (µPAC) for the analysis of permethylated glycans. Analyst 146 10.1039/d1an00643f (2021). Berg, H. E. et al. Micro-pillar array columns (µPAC): An efficient tool for comparing tissue and cultured cells of glioblastoma. J. Chromatogr. Open. 2 , 100047 (2025). B, V. & G, D. J, O. d. B. Column-Only Band Broadening in a Porous Shell Radially Elongated Pillar Array Column. Analytical chemistry 96, (2024). 10.1021/acs.analchem.3c05756 J, D. & P, L. C, B., A, S. Isolation Window Optimization of Data-Independent Acquisition Using Predicted Libraries for Deep and Accurate Proteome Profiling. Analytical chemistry 92, (2020). 10.1021/acs.analchem.0c00994 TS, B. & JV, O. Offline High pH Reversed-Phase Peptide Fractionation for Deep Phosphoproteome Coverage. Methods Mol. biology (Clifton N J) . 1355 10.1007/978-1-4939-3049-4_12 (2016). F, Y., DG, Y. S., RD, S. & C. & High-pH reversed-phase chromatography with fraction concatenation for 2D proteomic analysis. Expert Rev. Proteomics . 9 10.1586/epr.12.15 (2012). D, Y. et al. Separation Orthogonality in Liquid Chromatography-Mass Spectrometry for Proteomic Applications: Comparison of 16 Different Two-Dimensional Combinations. Anal. Chem. 92 10.1021/acs.analchem.9b05407 (2020). J, P., GN, M. A. & JA, S. BC, C. A gas phase fractionation acquisition scheme integrating ion mobility for rapid diaPASEF library generation. Proteomics 23 10.1002/pmic.202200038 (2023). RB, K., JC, Y. & YJ, C. Advances in data-independent acquisition mass spectrometry towards comprehensive digital proteome landscape. Mass Spectrom. Rev. 42 10.1002/mas.21781 (2023). C, G. et al. Extensive and Accurate Benchmarking of DIA Acquisition Methods and Software Tools Using a Complex Proteomic Standard. J. Proteome Res. 20 10.1021/acs.jproteome.1c00490 (2021). CA, S. et al. Slow-twitch fiber proportion in skeletal muscle correlates with insulin responsiveness. J. Clin. Endocrinol. Metab. 98 10.1210/jc.2012-3876 (2013). PH, A. et al. Human muscle fiber type-specific insulin signaling: impact of obesity and type 2 diabetes. Diabetes 64 10.2337/db14-0590 (2015). M, M. et al. Protein profile of fiber types in human skeletal muscle: a single-fiber proteomics study. Skelet. Muscle . 11 10.1186/s13395-021-00279-0 (2021). C, L. et al. Repeated sprint training in hypoxia induces specific skeletal muscle adaptations through S100A protein signaling. FASEB journal: official publication Federation Am. Soc. Experimental Biology . 38 10.1096/fj.202302084RR (2024). P, S., KF, P., GI, S. & JW, K. Mitochondrial Dysfunction, Insulin Resistance, and Potential Genetic Implications. Endocrinology 161, (2020). 10.1210/endocr/bqaa017 GN, R., AL, C. & TM, C. KS, N. Altered mitochondrial function in insulin-deficient and insulin-resistant states. J. Clin. Investig. 128 10.1172/JCI120843 (2018). CB, N. Interplay between lipids and branched-chain amino acids in development of insulin resistance. Cell Metabol. 15 10.1016/j.cmet.2012.01.024 (2012). H, Z. et al. Branched-chain amino acid supplementation impairs insulin sensitivity and promotes lipogenesis during exercise in diet-induced obese mice. Obes. (Silver Spring Md) . 30 10.1002/oby.23394 (2022). C, W., WJ, G. X. S. A. T., CT, L. & F. & Transcriptional profiles of type 2 diabetes in human skeletal muscle reveal insulin resistance, metabolic defects, apoptosis, and molecular signatures of immune activation in response to infections. Biochem. Biophys. Res. Commun. 482 10.1016/j.bbrc.2016.11.055 (2017). RA, A. & V, S. Role of complement in diabetes. Mol. Immunol. 114 10.1016/j.molimm.2019.07.031 (2019). H, E. Influence of Obesity and Type 2 Diabetes on Calcium Handling by Skeletal Muscle: Spotlight on the Sarcoplasmic Reticulum and Mitochondria. Front. Physiol. 12 10.3389/fphys.2021.758316 (2021). G, S., GR, S. V., H, Z. & M., V, G. & Altered calcium homeostasis is correlated with abnormalities of fasting serum glucose, insulin resistance, and beta-cell function in the Newfoundland population. Diabetes 54 10.2337/diabetes.54.11.3336 (2005). H, D. et al. Advanced glycation end products induce skeletal muscle atrophy and insulin resistance via activating ROS-mediated ER stress PERK/FOXO1 signaling. Am. J. Physiol. Endocrinol. Metab. 324 10.1152/ajpendo.00218.2022 (2023). Z, L. et al. Vitamin B6 Prevents Endothelial Dysfunction, Insulin Resistance, and Hepatic Lipid Accumulation in Apoe (-/-) Mice Fed with High-Fat Diet. Journal of diabetes research (2016). 10.1155/2016/1748065 (2016). E, M. & F, V. Vitamin B6 and Diabetes: Relationship and Molecular Mechanisms. Int. J. Mol. Sci. 21 10.3390/ijms21103669 (2020). F, C. et al. Exercise activates autophagy and regulates endoplasmic reticulum stress in muscle of high-fat diet mice to alleviate insulin resistance. Biochem. Biophys. Res. Commun. 601 10.1016/j.bbrc.2022.02.058 (2022). M, F., F, F. & E, H., P, F. & New insights into ER stress-induced insulin resistance. Trends Endocrinol. Metab. 23 10.1016/j.tem.2012.06.003 (2012). Y. The PRIDE database resources in 2022: a hub for mass spectrometry-based proteomics evidences. Nucleic Acids Res. 50 10.1093/nar/gkab1038 (2022). Additional Declarations No competing interests reported. Supplementary Files SciRepSupplement228032025.xlsx SciRepSupplement128032025.pdf Cite Share Download PDF Status: Under Review Version 1 posted Editorial decision: Revision requested 13 Jun, 2025 Reviews received at journal 12 Jun, 2025 Reviews received at journal 10 Jun, 2025 Reviewers agreed at journal 07 Jun, 2025 Reviewers agreed at journal 04 Jun, 2025 Reviewers invited by journal 02 Apr, 2025 Editor assigned by journal 02 Apr, 2025 Editor invited by journal 02 Apr, 2025 Submission checks completed at journal 31 Mar, 2025 First submitted to journal 28 Mar, 2025 You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-6331082","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Article","associatedPublications":[],"authors":[{"id":441362692,"identity":"3310a493-a59d-4e33-9b6e-b41d8aab678b","order_by":0,"name":"Anna Czajkowska","email":"","orcid":"","institution":"Medical University of Białystok","correspondingAuthor":false,"prefix":"","firstName":"Anna","middleName":"","lastName":"Czajkowska","suffix":""},{"id":441362693,"identity":"58749128-f2d5-4596-8e8a-5dd308dbba0e","order_by":1,"name":"Łukasz Szczerbiński","email":"","orcid":"","institution":"Medical University of Białystok","correspondingAuthor":false,"prefix":"","firstName":"Łukasz","middleName":"","lastName":"Szczerbiński","suffix":""},{"id":441362694,"identity":"9284f065-3231-481f-8e78-4e123ddcf096","order_by":2,"name":"Macin Czajkowski","email":"","orcid":"","institution":"Bialystok University of Technology","correspondingAuthor":false,"prefix":"","firstName":"Macin","middleName":"","lastName":"Czajkowski","suffix":""},{"id":441362695,"identity":"3344cd8f-afaa-4e0c-81c3-4df493e61057","order_by":3,"name":"Anna Citko-Rojewska","email":"","orcid":"","institution":"Medical University of Białystok","correspondingAuthor":false,"prefix":"","firstName":"Anna","middleName":"","lastName":"Citko-Rojewska","suffix":""},{"id":441362696,"identity":"0aefc007-ccc6-480d-ac33-f8fc33cb088b","order_by":4,"name":"Paulina Konopka","email":"","orcid":"","institution":"Medical University of Białystok","correspondingAuthor":false,"prefix":"","firstName":"Paulina","middleName":"","lastName":"Konopka","suffix":""},{"id":441362697,"identity":"09d753a9-e567-43de-aae8-e100856da9f7","order_by":5,"name":"Agnieszka Urszula Błachnio-Zabielska","email":"","orcid":"","institution":"Medical University of Białystok","correspondingAuthor":false,"prefix":"","firstName":"Agnieszka","middleName":"Urszula","lastName":"Błachnio-Zabielska","suffix":""},{"id":441362699,"identity":"f3b27248-d8e9-4f58-9bae-07ee6b3d573a","order_by":6,"name":"Adam Krętowski","email":"","orcid":"","institution":"Medical University of Białystok","correspondingAuthor":false,"prefix":"","firstName":"Adam","middleName":"","lastName":"Krętowski","suffix":""},{"id":441362701,"identity":"4a105a2b-6e3e-42d1-a47f-987126e82940","order_by":7,"name":"Piotr Zabielski","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAABC0lEQVRIiWNgGAWjYDACZihtwMzAeIDhgIQcAwMPkMsGxOw4tTA2QLUwgLQYI7Qw49DCANPCANbCkNhASIs5O/PxxxUMdfJAxoGDP85YpG+4dvYAw4eywwzmOLRYNrMlNp5hYDPc2cyWcJjnhkTuhtt5CYwzzh0GSmHXYnCYx7AR6BbGDYd5DA4zfABpyTFg5m07DJTCq0XCfsNh/g8Hf3yQSDcAaflLWItBItAWhgNAhyWAtTDi0QLyy8wGg4RkoF+A2s9IGM4E+uVgz7l0Hlx+Mec/fOBjQ0Wd7Xb+ww8f/jhWJ893O/fggx9l1nLm7A3YHYZEIsABIOZBF0TVQqrUKBgFo2AUjCwAAN5YX1Gjn2A9AAAAAElFTkSuQmCC","orcid":"","institution":"Medical University of Białystok","correspondingAuthor":true,"prefix":"","firstName":"Piotr","middleName":"","lastName":"Zabielski","suffix":""}],"badges":[],"createdAt":"2025-03-28 23:23:11","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-6331082/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-6331082/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":81178182,"identity":"f64ff72c-5f13-4944-9287-983da5bba446","added_by":"auto","created_at":"2025-04-23 06:52:13","extension":"jpg","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":314249,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eSchematics of the study design\u003c/strong\u003e. Proteins from post-training, vastus lateralis muscle LN2 pulverizates were extracted with SDC-assisted lysis and sonication. Approx. 20µg of protein from each biopsy was combined into study-wide mixed sample. This sample – after digestion – was used to create ion-chromatogram libraries through High-pH fractionation with fraction concatenation (HpH, 6 final fractions), gas-phase fractionation approach (GPF, 6 fractions) or repetitive injections (libraryless DirectDIA\u003csup\u003eTM\u003c/sup\u003e\u0026nbsp; approach, 6 repetitive injections). Assessed library combinations included data collected in DDA mode, STW mode (staggered windows DIA), or both. Subsequently, to assess performance of approaches in protein identification and quantitation, three best-performing library combinations were used to analyze data from mixed muscle samples collected in STW mode. Finally, the best-performing libraries were utilized for the IGT vs. NGT differential protein expression analysis across the entire set of experimental samples.\u003c/p\u003e","description":"","filename":"fig1.jpg","url":"https://assets-eu.researchsquare.com/files/rs-6331082/v1/8794447b5c30723f38569804.jpg"},{"id":81178184,"identity":"e5167ff1-48ae-4095-aa50-839e710a7faa","added_by":"auto","created_at":"2025-04-23 06:52:13","extension":"jpg","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":427632,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eComparison of the fractionation-based and computational-based libraries selected for the further evaluation. \u003c/strong\u003eVenn diagrams of library overlap at precursor \u003cstrong\u003e(Panel A)\u003c/strong\u003e, peptide \u003cstrong\u003e(Panel B)\u003c/strong\u003eand protein \u003cstrong\u003e(Panel C)\u003c/strong\u003e level; normalized histogram of precursor (peptide-spectrum match, PSM) intensity distribution \u003cstrong\u003e(Panel D)\u003c/strong\u003e; radar plot of the peptide physicochemical properties \u003cstrong\u003e(Panel E)\u003c/strong\u003e; pathway enrichment analysis of selected diabetes-related molecular pathways \u003cstrong\u003e(Panel F)\u003c/strong\u003e; basic statistics of molecular pathways identified within a given ion-chromatogram library (table insert). Panel F bubble size represents significance of enrichment of a given pathway expressed as –Log10 p-value, whereas Y-axis represents the total number of identified molecular members (proteins) of a given pathway.\u003c/p\u003e","description":"","filename":"fig2.jpg","url":"https://assets-eu.researchsquare.com/files/rs-6331082/v1/4982081d3908e5d1ae92df46.jpg"},{"id":81177106,"identity":"4761fb8a-e2cc-421e-a620-3e3f0c8255eb","added_by":"auto","created_at":"2025-04-23 06:36:13","extension":"jpg","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":511552,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eComparison of the results of the skeletal muscle mixed sample analysis (n=3) performed with the use of selected approaches. \u0026nbsp;\u003c/strong\u003eVenn diagrams of muscle proteins at precursor \u003cstrong\u003e(Panel A)\u003c/strong\u003e, peptide \u003cstrong\u003e(Panel B)\u003c/strong\u003e, and protein \u003cstrong\u003e(Panel C)\u003c/strong\u003e levels; radar plot of the peptide physicochemical properties \u003cstrong\u003e(Panel D)\u003c/strong\u003e; assay precision at precursor, peptide, and protein levels \u003cstrong\u003e(Panel E)\u003c/strong\u003e; protein rank distribution plot \u003cstrong\u003e(Panel F)\u003c/strong\u003e; correlation plot at precursor (PSM), peptide, and protein level between the selected approaches (all correlations at p\u0026lt;0.00001). \u003cstrong\u003eSupplement 1 Figure S8\u003c/strong\u003e presents bubble plot of pathway enrichment analysis and basic statistics of molecular pathways identified by a given approach in the skeletal muscle mixed samples.\u003c/p\u003e","description":"","filename":"fig3.jpg","url":"https://assets-eu.researchsquare.com/files/rs-6331082/v1/224a52f045a71021a2390e5a.jpg"},{"id":81177758,"identity":"047211bd-734a-4e20-8cd1-30568323bb8a","added_by":"auto","created_at":"2025-04-23 06:44:13","extension":"jpg","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":588880,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eApplication of the selected approaches for the analysis of post-exercise muscle proteome changes in glucose-intolerant subjects (post-exercise IGT vs NGT). \u003c/strong\u003eStudy-wide\u003cstrong\u003e \u003c/strong\u003epathway enrichment analysis of selected diabetes-related molecular pathways \u003cstrong\u003e(Panel A)\u003c/strong\u003e,\u003cstrong\u003e \u003c/strong\u003ewith basic statistics of significantly enriched pathways (table insert); comparison of the results of pathway enrichment analysis (z-scores of regulation) showing selection of pathways relevant to preservation of healthy muscle physiology \u003cstrong\u003e(Panel B)\u003c/strong\u003e; comparison of significantly affected skeletal muscle biological functions and potential muscular disorders \u003cstrong\u003e(Panel C)\u003c/strong\u003e.\u003c/p\u003e","description":"","filename":"fig4.jpg","url":"https://assets-eu.researchsquare.com/files/rs-6331082/v1/a91f4f1ba95810769e1a7dc2.jpg"},{"id":81177112,"identity":"8aaffc66-2a21-4017-8151-896f318725b4","added_by":"auto","created_at":"2025-04-23 06:36:13","extension":"jpg","order_by":5,"title":"Figure 5","display":"","copyAsset":false,"role":"figure","size":258650,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eProtein functional clustering of IGT vs NGT dataset analyzed with HPH/DDA-H ion-chromatogram library.\u003c/strong\u003e Figure shows functional clustering within 140 significantly affected proteins (orphans excluded) identified by STRING network. Analysis revealed 4 major interconnected clusters which combine proteins involved in regulation of translation and RNA processing \u003cstrong\u003e(Cluster I)\u003c/strong\u003e, protein processing in ER (endoplasmic reticulum) \u003cstrong\u003e(Cluster II)\u003c/strong\u003e, ER to Golgi vesicular protein transport \u003cstrong\u003e(Cluster III)\u003c/strong\u003e and structural constituents of skeletal muscle sarcomere \u003cstrong\u003e(Cluster IV)\u003c/strong\u003e. Two additional clusters aggregated proteins involved in mitochondrial fatty acids metabolism \u003cstrong\u003e(Cluster V) \u003c/strong\u003eand endosomal protein processing \u003cstrong\u003e(Cluster VI)\u003c/strong\u003e. Thick dashed lines denote areas occupied by clusters involved in similar biological processes. Shading of halo depicts significance of differential protein expression between NGT and IGT (expressed as –log\u003csub\u003e10\u003c/sub\u003e(p-value)). Line thickness indicates the strength of node interaction (edge confidence) between the members of a given cluster. Thin dashed lines represent node inter-cluster functional connections. Full results of STRING analysis are presented in \u003cstrong\u003eSupplement 2 Table S9.\u003c/strong\u003e\u003c/p\u003e","description":"","filename":"fig5.jpg","url":"https://assets-eu.researchsquare.com/files/rs-6331082/v1/855214e66e652f1f6723fb30.jpg"},{"id":81177760,"identity":"0f8fccab-f011-42b7-a017-ef41b9717656","added_by":"auto","created_at":"2025-04-23 06:44:13","extension":"jpg","order_by":6,"title":"Figure 6","display":"","copyAsset":false,"role":"figure","size":214996,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eResults of protein functional clustering from IGT vs NGT dataset analyzed with GPF/STW-H ion-chromatogram library. \u003c/strong\u003eFigure shows functional clustering within 130 significantly affected proteins (orphans excluded) identified by STRING network. Analysis revealed 3 major interconnected clusters which aggregate proteins involved in RNA splicing, lipoprotein processing and lipid uptake, and pentose phosphate pathway \u003cstrong\u003e(Cluster I)\u003c/strong\u003e, protein synthesis, RNA binding and translation, peptide degradation and cellular proliferation \u003cstrong\u003e(Cluster II, with 2 subclusters) \u003c/strong\u003eand skeletal muscle contraction and myocyte-extracellular matrix interaction \u003cstrong\u003e(Cluster III)\u003c/strong\u003e.\u003cstrong\u003e \u003c/strong\u003eAdditional clusters combine proteins involved in branched-chain amino acid metabolism and skeletal muscle carnosine metabolism \u003cstrong\u003e(Cluster IV, with 1 subcluster)\u003c/strong\u003e, and pyridoxal phosphate (vitamin B\u003csub\u003e6\u003c/sub\u003e) synthesis and transport \u003cstrong\u003e(Cluster V)\u003c/strong\u003e. Thick dashed lines denote areas occupied by clusters involved in similar biological processes. Shading of halo depicts significance of differential protein change between NGT and IGT (expressed as -log\u003csub\u003e10\u003c/sub\u003e(p-value)). Line thickness indicates the strength of node interaction (edge confidence) between the members of a given cluster. Thin dashed lines represent node inter-cluster functional connections. Full results of STRING analysis are presented in \u003cstrong\u003eSupplement 2 Table 10.\u003c/strong\u003e\u003c/p\u003e","description":"","filename":"fig6.jpg","url":"https://assets-eu.researchsquare.com/files/rs-6331082/v1/b51e91b26aa3df0b404917cd.jpg"},{"id":81179000,"identity":"760bc873-6c9b-4641-980b-a5ac9f8a611a","added_by":"auto","created_at":"2025-04-23 07:00:14","extension":"jpg","order_by":7,"title":"Figure 7","display":"","copyAsset":false,"role":"figure","size":117559,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eProtein functional clustering of NGT vs IGT dataset analyzed with DirectDIA\u003c/strong\u003e\u003csup\u003e\u003cstrong\u003eTM\u003c/strong\u003e\u003c/sup\u003e\u003cstrong\u003e approach.\u003c/strong\u003e Figure shows functional clustering within 64 significantly affected proteins (orphans excluded) identified by STRING network. A total of 5 non-connected clusters aggregate proteins involved in muscle contraction and myocyte-intracellular matrix interaction \u003cstrong\u003e(Cluster I)\u003c/strong\u003e, protein synthesis and proteasomal degradation \u003cstrong\u003e(Cluster II)\u003c/strong\u003e, intra- and extracellular Ca\u003csup\u003e2+ \u003c/sup\u003ecalcium binding \u003cstrong\u003e(Cluster III)\u003c/strong\u003e, protein PTM (glycation) \u003cstrong\u003e(Cluster IV)\u003c/strong\u003e, and mitochondrial medium and short-chain acyl-CoA synthesis \u003cstrong\u003e(Cluster V)\u003c/strong\u003e. Thick dashed lines denote areas occupied by clusters involved in similar biological processes. Shading of halo depicts significance of differential protein expression between NGT and IGT (expressed as –log\u003csub\u003e10\u003c/sub\u003e(p-value)). Line thickness indicates the strength of node interaction (edge confidence) between the members of a given cluster. Full results of STRING analysis are presented in \u003cstrong\u003eSupplement 2 Table 11\u003c/strong\u003e.\u003c/p\u003e","description":"","filename":"fig7.jpg","url":"https://assets-eu.researchsquare.com/files/rs-6331082/v1/d77d56f802a0490614612edc.jpg"},{"id":81179440,"identity":"5ed618d2-42b5-41e0-851e-fb02a1076cc9","added_by":"auto","created_at":"2025-04-23 07:08:16","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":4520639,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-6331082/v1/803f39b3-0b13-43eb-b0af-6413488295c3.pdf"},{"id":81177115,"identity":"66c3f99a-61b2-4ccb-91f5-3e3d1c1fb0c7","added_by":"auto","created_at":"2025-04-23 06:36:13","extension":"xlsx","order_by":0,"title":"","display":"","copyAsset":false,"role":"supplement","size":587320,"visible":true,"origin":"","legend":"","description":"","filename":"SciRepSupplement228032025.xlsx","url":"https://assets-eu.researchsquare.com/files/rs-6331082/v1/2125b958d53e41a32eb923c4.xlsx"},{"id":81177121,"identity":"47c4731e-2ac8-4901-84d1-4edfb48b29f4","added_by":"auto","created_at":"2025-04-23 06:36:14","extension":"pdf","order_by":1,"title":"","display":"","copyAsset":false,"role":"supplement","size":7839316,"visible":true,"origin":"","legend":"","description":"","filename":"SciRepSupplement128032025.pdf","url":"https://assets-eu.researchsquare.com/files/rs-6331082/v1/436e75da2d744f73aaeea9c2.pdf"}],"financialInterests":"No competing interests reported.","formattedTitle":"Ion-chromatogram libraries assembly in DIA proteomic analysis of post-exercise skeletal muscle in prediabetic subjects","fulltext":[{"header":"1.\tIntroduction","content":"\u003cp\u003eType 2 diabetes mellitus (T2DM) and its comorbidities such as cardiovascular-related disorders, hyperlipidemia, obesity, nephropathy, neuropathy and decrease in cognitive function and dementia become a major health-related burden worldwide \u003csup\u003e1,2\u003c/sup\u003e. Progression towards T2DM is connected with increasing insulin resistance of major insulin-sensitive tissues (muscle, adipose tissue, liver) with subsequent induction of pre-diabetic state. This usually asymptomatic, obscure metabolic abnormality can manifest itself as isolated impaired fasting glucose (IFG \u0026ndash; fasting blood glucose between 100mg/dl and 125mg/dl) or impaired glucose tolerance (IGT \u0026ndash; 2 hour blood glucose after 75g oral glucose challenge between 140mg/dl and 199mg/dl) and elevated blood insulin \u003csup\u003e3\u003c/sup\u003e. As prediabetes is often overlooked, pre-diabetes progresses toward symptomatic T2DM after\u0026nbsp;b-cell failure, when pancreatic islets are unable to compensate for increasing systemic insulin demands. This crucial watershed event marks an irreversible stage of T2DM pathogenesis\u0026nbsp;\u003csup\u003e4,5\u003c/sup\u003e. Currently, due to development of various pharmacotherapies (SGLT2 inhibitors, DPP4 inhibitors, GLP1 analogs and receptor agonists, insulin sensitizers and long-acting insulin analogs) case-tailored T2DM management is increasingly successful, leading to both a decrease in hyperglycemia and T2DM comorbidities\u0026nbsp;\u003csup\u003e6,7\u003c/sup\u003e. Contrary to fully developed T2DM,\u0026nbsp;b-cell dysfunction and systemic insulin resistance in prediabetes is reversible and can be successfully managed by approaches such as bariatric surgery, pharmacotherapy, lifestyle changes, physical exercise and their combinations\u0026nbsp;\u003csup\u003e8-11\u003c/sup\u003e. From the above, physical exercise appears as an easily accessible, universal remedy in both the prediabetes prevention, augmentation of its reversal or inhibition of its progression towards T2DM\u0026nbsp;\u003csup\u003e10,12,13\u003c/sup\u003e. Moreover, physical exercise has shown beneficial effects in the prevention of other metabolic-syndrome connected comorbidities such as cognitive impairment and dementia\u0026nbsp;\u003csup\u003e14-16\u003c/sup\u003e.\u003c/p\u003e\n\u003cp\u003eMetabolic health of muscle tissue is of great importance for both the physical fitness and whole-body energy metabolism. As a major tissue in both insulin-stimulated and insulin-independent (exercise-evoked) uptake and metabolism of plasma glucose, its role in prediabetes prevention and its reversal cannot be overlooked. Prediabetes was shown to induce detrimental effects in both the metabolic and contractile muscle function \u003csup\u003e17-20\u003c/sup\u003e. Crucially, \u0026nbsp; exercise in different forms and modalities was shown to improve both metabolic and functional abnormalities observed in prediabetes \u003csup\u003e21,22\u003c/sup\u003e. Taking all of the above into account, analysis of the impact of exercise on the proteome of prediabetic skeletal muscle is crucial for the elucidation of its beneficial effects. Moreover, the subject of equal importance is the identification of persistent negative effects of prediabetes, which are not fully corrected by the physical activity. However, proteomic analysis of skeletal muscle is burdensome due to several difficulties arising from the biological and molecular characteristics. At the macroscopic level these include site-dependent variations in connective tissue content, extracellular matrix deposition and adipose tissue distribution, which \u0026ndash; despite careful sampling and post-biopsy cleanup \u0026ndash; induce significant variation in sample composition. Moreover, developed striated muscle is composed mainly of chains of biological polymers involved in muscle contraction, such as myosin, actin, tropomyosin and titin chains \u003csup\u003e23,24\u003c/sup\u003e. This creates so-called \u0026ldquo;tip of the iceberg effect\u0026rdquo;, similar to that of plasma proteome, where a number of proteins are greatly overrepresented in non-depleted samples, which subsequently masks low-abundant proteins \u003csup\u003e25,26\u003c/sup\u003e. Thus, methods tailored for proteomic analysis of skeletal muscle cannot be directly derived from those developed using easily accessible matrices, such as cell-culture material. Cultured cells are inherently less demanding and yield high proteome coverage levels, presenting best-case scenarios in proteomic sample analysis.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eRobust methods for proteomic analysis of clinical skeletal muscle samples need to address the specificity of this tissue at both the LC/MS/MS level and subsequent bioinformatic analysis. At the HPLC peptide separation level, it requires the use of traps and columns resistant to clogging by long biological polymers, flow-reversible for easy cleanup and restoration of its initial performance, with robust retention time stability after hundreds of samples \u003csup\u003e27,28\u003c/sup\u003e. Ideally, with low operating pressure and long separation length and adequate resolution to address \u0026ldquo;tip-of the iceberg\u0026rdquo; effect of non-depleted samples. At the MS level, it dictates the use of data-independent acquisition (DIA), ideally supplemented with retention time calibration peptides, which is better suited for clinical applications due to its reproducibility and non-stochastic nature of identification, unlike data-dependent acquisition (DDA) \u003csup\u003e29,30\u003c/sup\u003e. Finally, employment of DIA requires assembly of appropriate ion-chromatogram libraries, or selection of libraryless, pure bioinformatic identification and quantitation pipeline \u003csup\u003e31-33\u003c/sup\u003e.\u003c/p\u003e\n\u003cp\u003eTo address above challenges, we constructed a pipeline tailored for proteomic analysis of clinical skeletal muscle biopsies. We employed\u0026nbsp;mPAC semiconductor-technology chromatography columns and traps, retention-time calibration peptides and staggered-windows DIA analysis to address reproducibility issues inherent to challenging sample matrices. As employment of different approaches to ion-chromatogram library creation can lead to distinct biological findings, \u0026nbsp;we compared various ion-chromatogram libraries based on off-line high-pH fractionation (HpH), \u0026nbsp;gas-phase fractionation (GPF) and libraryless DirectDIA\u003csup\u003eTM\u003c/sup\u003e approach. \u0026nbsp;Finally, we performed proteomic analysis of muscle from normoglycemic (NGT) and prediabetic (IGT) subjects after 3 months of supervised, mixed-mode exercise\u0026nbsp;\u003c/p\u003e"},{"header":"2. Materials and Methods","content":"\u003cp\u003e\u003cstrong\u003e2.1 Study participants and sample collection:\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u0026nbsp;The muscle biopsy samples were collected from participants of the “Bialystok Exercise Study in Diabetes (BESD)”, conducted by the Department of Endocrinology, Diabetology and Internal Medicine and Clinical Research Centre of the Medical University of Bialystok \u003csup\u003e22,34\u003c/sup\u003e. The study and its procedures were accepted by Ethics Committee of the Medical University of Bialystok (No. R-I-002/469/2014). Each participant signed informed consent before participation and subsequent sample collection. Briefly, the study cohort involved physically inactive men with various degrees of dysglycemia, with a BMI of 25-35 kg/m\u003csup\u003e2\u003c/sup\u003e and leading a sedentary lifestyle. All participants underwent three-month exercise program that included supervised training sessions at a local fitness center. Studied patients participated in a 3-month exercise intervention consisting of mixed training, aerobic and strength exercises, performed for 85 minutes 3 times a week, totaling in 36 sessions over 12 weeks \u003csup\u003e22,34\u003c/sup\u003e. The routine started with a 15-minute warm-up, followed by 40 minutes of strength exercise of major muscle groups (60-75% of single repetition maximum, adjusted weekly) and 30 minutes of moderate-intensity endurance exercise (60–70% of individual VO\u003csub\u003e2\u003c/sub\u003emax), all supervised by an exercise technician. The myWellness system (Technogym, Cesena, Italy) recorded each session for consistency. Dietary supervision aimed to eliminate diet-related variables. For the assessment of proposed methodology, we selected only the post-exercise bioptates from non-diabetic, normoglycemic, with normal glucose tolerance (NGT group; n=13, fasting plasma glucose (FPG) \u0026lt;100 mg/dL and 2-h glucose during OGTT (2h-OGTT-GLU) \u0026lt;140 mg/dL) and prediabetic, impaired glucose tolerance group (IGT group; n=11; FPG \u0026lt;100 mg/dL, 2h-OGTT-GLU 140–199 mg/dL). Analysis was performed on 2 independent bioptates from each participant, yielding a total of 26 NGT and 22 IGT samples (total of 48 individual samples). Post-exercise anthropometric characteristics of NGT and IGT groups are presented in \u003cstrong\u003eSupplement 1 (Table S1)\u003c/strong\u003e.The vastus lateralis muscle (VL) biopsies were obtained from fasted participants 48 hours after the last exercise session with the use of a percutaneous suction needle \u003csup\u003e35\u003c/sup\u003e. Excess blood, connective tissue and fat was removed after visual inspection immediately after collection. All samples were snap-frozen and stored in LN\u003csub\u003e2\u003c/sub\u003e until further analysis.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e2.2 Experimental Design:\u003c/strong\u003e\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eExperimental design of the study is presented in \u003cstrong\u003eFigure 1\u003c/strong\u003e.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e2.3 Sample preparation:\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eProtein extraction and digestion was performed with the use of sodium deoxycholate-assisted sample preparation with phase transfer SDC extraction and lipid depletion technique according to Leon et al. \u003csup\u003e36\u003c/sup\u003e. Briefly, approx. 25-30mg of lateral thigh muscle biopsy was pulverized in LN\u003csub\u003e2\u003c/sub\u003e, suspended in tissue lysis buffer (1:10 w/v; 50mM ammonium bicarbonate (ABC), 5% sodium deoxycholate (SDC), 5mM TCEP) and sonicated for 2 x 30 sec. on ice at 50% power (Sonics\u0026amp;Materials VCX-130 Ultrasonic Processor). Subsequently, samples were denaturated by heating for 30 min. at 60°C, alkylated with iodoacetamide (IAA) for 30 min. at room temperature (IAA, 1M in ABC buffer, 15mM final concentration) and centrifuged for 5min. at \u0026nbsp;14000g to sediment tissue debris. Protein concentration was measured with BCA assay (reducing-agent compatible, Thermo Scientific) and protein content was normalized to 2.5 μg/μl with 5% SDC in 50mM ABC buffer. Digestion-ready samples were stored in-80°C until further analysis. For sample digestion, 10μl of sample equivalent to 25μg of protein, was diluted to 90μl with 50mM ABC. Trypsin/LysC mix (2.5μl from 0.4μg/μl stock in 50mM acetic acid) was added at a ratio of 1:25 enzyme/protein (1:50 ratio for each of the enzymes). Samples were incubated at 37°C for 12h with shaking (600 rpm) in Thermoblock, and subsequently cooled and kept at 4°C until next working day. After addition of TFA to 0.5% concentration and brief vortexing, precipitated SDC and lipids were extracted 3 times with 100μl of ethyl acetate. Each time samples were vortexed and centrifuged for 5 min at 14000g. Residual ethyl acetate was evaporated by heating the open tubes at 60°C for 30 minutes with gentle shaking. Subsequently, 10 injection equivalents of iRT peptides (Biognosys) were added to each sample. Additionally, approx. \u0026nbsp;20μg of protein was taken from each homogenized biopsy to create study-wide mixed sample. After protein digestion, this mixed sample was used for offline high-pH fractionation (HpH), gas-phase fractionation (GPF) and for repetitive \u0026nbsp; injections performed at the library selection stage. All injection-ready samples were stored in -80°C until further analysis.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e2.4 LC/MS/MS analysis:\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe same optimized chromatographic conditions were used for both data-dependent (DDA) and data-independent (DIA) analysis. Detailed description and results of optimization of chromatographic gradient, column flow and sample loading for µPAC™ pillar array trap and column are presented in\u003cstrong\u003e\u0026nbsp;Supplement 1 (Table S2-S4).\u0026nbsp;\u003c/strong\u003eAll optimization steps were performed on HeLa digest standard (Thermo Scientific) run in triplicate in DDA mode.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eLC/MS/MS analysis was performed on Q-Exactive MS equipped with IonFlex II source and silica glass emitter (10μm or 20μm I.D. PicoTip, flow-rate dependent; NewObjective) working at 1.75kV in +ESI mode. Thermo Scientific 3500RSLC nanoLC configured for trap-elute was used in all chromatographic separations. For standard runs, samples were loaded at 750ng of peptides onto µPAC™ nano pillar array trap at 10μl/min (2% ACN, 0.2% TFA) and resolved on 50cm µPAC™ pillar array column (channel length 500mm, channel width 315μm, pillar height 18μm, pillar diameter 5μm, interpillar distance 2.5μm) at 300nl/min using 115min multi step reverse-phase binary gradient (A – 0.2% FA in H\u003csub\u003e2\u003c/sub\u003eO, LC/MS; B – 0.2% FA in 90% acetonitrile, LC/MS; 5%B-4min, 30%B-71min, 45%B-94min, 90%B-95min, 90%B-103min, 5%B-104min, 5%B-115min). Total sample run time excluding autosampler draw equaled 120min. Daily routine\u0026nbsp;started with mass calibration, nanoLC solvent purge and column equilibration. Each batch of 11 samples was preceded by iRT-spiked QC sample (300ng of HeLa digest standard, Thermo Scientific) analyzed with 55 min gradient and DDA acquisition. Study-wide analysis of sample-spiked iRT peptide retention time variation, FWHM and mass accuracy and total ion chromatograms of all experimental samples and respective mass accuracy at MS and MS/MS level is presented in \u003cstrong\u003eSupplement 1 (Figure S1 and S2)\u003c/strong\u003e.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e2.5 Data-dependent (DDA) acquisition:\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eDDA analysis of eluting peptides for both the optimization of chromatographic conditions and library generation was performed using Top 20 method with following settings: MS1 resolution of 70k, ACG target 3e\u003csup\u003e6\u003c/sup\u003e, max. IT 50ms, scan range 385-1015m/z; MS2 resolution 17k, ACG target 2e\u003csup\u003e5\u003c/sup\u003e, max. IT 55ms, isolation 2m/z with 0.4mz offset to include most of isotopologues in case of multiply-charged peptides. Top 20 peptides (excluding isotopes) were selected for fragmentation with normalized collision energy (NCE) lowered to 27, \u0026nbsp;if following criteria were fulfilled: charged state 2-5, preferred peptide match, min. ACG 6e\u003csup\u003e3\u0026nbsp;\u003c/sup\u003eand intensity threshold 1.1e\u003csup\u003e5\u003c/sup\u003e, with dynamic exclusion set at 15sec.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e2.6 Data-independent (DIA) acquisition and gas-phase fractionation (GPF):\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eDIA method for library generation and sample analysis (referred as 24m/z STW in the manuscript) was designed on the basis of staggered windows approach by Searle, Pino and Amodei \u003csup\u003e31,37-39\u003c/sup\u003e. Method consisted of 52 staggered, 24m/z wide windows (12m/z overlap), covering 400-1000 m/z (referred as 24SW method) with MS1 scan covering 385-1015m/z performed after 26 windows \u003cstrong\u003e(Supplement 1, Figure S3; Supplement 2, Table S1)\u003c/strong\u003e. Mass range was selected on the basis of distribution of peptide m/z values from DDA run of mixed muscle sample. Window edges were placed in peptide forbidden zones to account for non-rectangular characteristics of Q-Exactive quadrupole mass isolation. Window placement was designed with Skyline. MS1 resolution was set at 35k, ACG at 1e\u003csup\u003e6\u003c/sup\u003e and max. IT of 60ms. Windowed MS2 was performed at 17k resolution, with ACG at 1e\u003csup\u003e6\u003c/sup\u003e, max. IT of 60ms, MSX count of 1 with isochronous IT set to ON and 27 NCE at default charge state of 3.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eFor gas-phase fractionation (GPF) with staggered windows approach (4 m/z GPF-STW), 750ng of\u0026nbsp;peptides was resolved for each of the 6 gas-phase fractions. Gas-phase fractionation consisted of 6 overlapping segments 102m/z wide (2m/z overlap) jointly covering 400-1000 m/z mass range \u003cstrong\u003e(Supplement 2, Table S2)\u003c/strong\u003e. Each gas-phase fraction consisted of 52 staggered, 4 m/z wide windows (2m/z overlap), with MS1 scan covering respective mas range performed after 26 windows and utilized identical MS1 and MS2 settings as DIA method described above. Results of GPF with staggered windows presented in \u003cstrong\u003eSupplement 1 (Figure S3).\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e2.7 Off-line high-pH peptide fractionation with fraction concatenation:\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eFor HpH fractionation with fraction concatenation (HpH), post-digestion peptides from mixed sample were evaporated to dryness using Labconco Vacuum Concentrator at 45°C under 0.05atm residual pressure. Subsequently, peptides were dissolved at 5μg/μl in 2% ACN in 100mM NH\u003csub\u003e4\u003c/sub\u003eOH with 0.001% Zwittergent 3-16 through vortexing and brief sonication. Up to 100μg (20μl) of peptides were loaded onto C18 Waters X-Bridge Peptide BEH Column (300Å, 3.5 µm, 1 mm X 150 mm) and resolved using Dionex 3500RSLC. Peptides were eluted at a flow rate of 100μl/min using a 45 min, reverse-phase binary gradient (A -2% ACN in 10mM NH\u003csub\u003e4\u003c/sub\u003eOH, B – 90% ACN in 10mM NH\u003csub\u003e4\u003c/sub\u003eOH; gradient 3.3%B for 2.5min, 25%B at 32.5min, 40%B at 35min, 90%B at 35.5min for 3.5min, 3.3%B at 42.5min for 2.5min) under UV monitoring at 214nm (peptide bond) and 280nm (tyrosine, tryptophan). Peptide fractions eluting for up to 42 min were collected every 140 seconds into 96 well 500µl LoBind plate (Eppendorf). Resulting 18 fractions were pooled to final 6 using concatenation scheme to enhance orthogonality of basic (high-pH) and acidic (low-pH) separations and equalize concentration and distribution of peptides. Combined peptide fractions were evaporated under vacuum and resuspended at 250ng/μl in sample loading buffer (0.5% TFA, 2% ACN, 0.001% Zwittergent). Spike of iRT peptides was added to each of concatenated samples. Subsequently, 1μg of peptides from each fraction was analyzed by both the DDA and DIA methods and resulting data was used to generate appropriate library combination. Results of HpH fractionation are presented in \u003cstrong\u003eSupplement 1 (Figure S5)\u003c/strong\u003e.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eMS cycle time of 3180ms was matched across all the methods and chromatographic runs to give approx. 8 points per peak at FWHM \u003cstrong\u003e(Supplement 1; Table S6)\u003c/strong\u003e. For all DDA, STW and GPF-STW methods, polysiloxane at 445.12003 m/z was used as lock mass for real-time calibration.\u0026nbsp;\u003c/p\u003e\n\u003col start=\"2\"\u003e\n \u003cli\u003e\u003cstrong\u003eIon-chromatogram libraries:\u003c/strong\u003e\u003c/li\u003e\n\u003c/ol\u003e\n\u003cp\u003eTo select optimal analysis approach, several versions of libraries were compiled. Each included combinations of appropriate fractions (GPF, HPH, collected in DDA and STW mode) and mixed sample runs (collected in STW mode). Spectronaut version 15.4.210913.50606-Rubin \u003csup\u003e40\u003c/sup\u003e, was used to search raw files with the build-in Pulsar engine against Homo sapiens Uniprot reference proteome FASTA file (UP000005640_9606, one sequence per protein). Pulsar standard settings were used, with following exceptions: digestion by Trypsin/LysC with P, carbamidometylation of C as fixed, acetylation of N-terminal, methylation of K or R, deamidation of Q or N, oxidation of M as variable modifications. Peptide length between 7 and 35 amino acids, max 5 variable modifications and max 2 missed cleavages per peptide. Spectra collected with staggered windows method were demultiplexed by build-in algorithm\u0026nbsp;\u003csup\u003e39\u003c/sup\u003e. Dynamic mass calibration was used to correct both MS1 and MS2 spectra. Library was constrained to include from 3 to 6 of most intense b or y type fragments (300mz – 1800m/z range) per peptide of at least 3AA in length. Universal 1% FDR cutoff was used at the level of precursors; peptides and proteins. Retention times and RT window widths were calibrated using deep learning-assisted iRT regression included in Spectronaut. Results of library generation step are presented in \u003cstrong\u003eSupplement 1 (Table S5)\u003c/strong\u003e.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eSubsequently, mixed sample runs collected with standard\u0026nbsp;24m/z STW method\u0026nbsp;(n=3)\u0026nbsp;were searched with different versions of libraries to select best-performing combinations. The results are presented in \u003cstrong\u003eSupplement 1 (Table S6)\u003c/strong\u003e.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e2.9 Data analysis:\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eBiognosys Spectronaut was used to extract ion chromatograms from MS1 and MS2 data based on both the maximum ion peak intensity using dynamic mass tolerance, dynamic iRT-corrected retention time window and global, MS2-level signal normalization (\u003cstrong\u003eSupplement 1 Figure S3\u003c/strong\u003e). Subsequently, data was searched with selected experiment-specific ion libraries. Q-value of 0.01 for both the precursors and peptides (corresponding to 1% FDR) was at the experiment level with a target-decoy approach using mutated sequences. For proteins Q-value as set at \u0026lt; 0.05 (5% FDR) and \u0026lt; 0.01 (1% FDR) at run-vise and experiment-vise levels, respectively. Quantification was performed at MS2 level, with Spectronaut bulid-in MaxLFQ algorithm. Within-sample peptide quantity was calculated as a mean of the peak areas of top 3 MS2 fragment ions, whereas protein quantity was expressed as a mean of top 3 peptides. For NGT and IGT group comparison, only the proteins present in at least 50% of the samples (Q-value percentile 0.5), with 2 unique peptides, presenting at least 50% expression difference (-0.585³\u0026nbsp;log\u003csub\u003e2\u003c/sub\u003e(fold change)\u0026nbsp;³\u0026nbsp;0.585) and FDR-adjusted p-value \u0026lt;0.05 (Q-value \u0026lt; 0.05, -log\u003csub\u003e10\u003c/sub\u003e(p-value) \u0026gt;1.13) were regarded as significant \u003csup\u003e40\u003c/sup\u003e. For molecular pathway and molecular interaction analysis, IGT/NGT protein Log2 expression rations and Q-values were uploaded to Quiagen IPA \u003csup\u003e41\u003c/sup\u003e. Both expression Q-value cutoffs were used to restrict dataset to only significantly affected proteins (as mentioned above). IPA Core Analysis was restricted to mammalian proteins (human, rat, mouse) present in skeletal muscle, using stringent filtering for both molecules and interactions. Z-scores for pathways, diseases and bio functions were calculated based on expression log2 ratios. Corresponding -log\u003csub\u003e10\u003c/sub\u003e(p-value) of overlap were calculated using right-tailed Fisher's exact test with and FDR correction. Compare analysis function in IPA was used to correlate IPA results from IGT/NGT comparison generated with the HpH, GPF and DirectDIA\u003csup\u003eTM\u0026nbsp;\u003c/sup\u003eapproach. Results of comparison analysis at molecular pathway, disease and bio-function and individual protein level are presented in \u003cstrong\u003eSupplement 2 (Table S3-S5)\u003c/strong\u003e. Detailed results of IPA analysis performed on samples quantified with respective library are presented in \u003cstrong\u003eSupplement 2 (Table S6-S8)\u003c/strong\u003e. Protein functional clustering was performed with String \u003csup\u003e42\u003c/sup\u003e. We conducted a standard gene set analysis of significantly affected proteins, using their gene names and ranked FDR-adjusted p-values (4 ranks) expressed as -log\u003csub\u003e10\u003c/sub\u003e(p-value). Analysis was performed with full STRING Homo sapiens database (text mining, experiments, databases, co‑expression, neighborhood, gene fusion, co‑occurrence), confidence of supporting data was used as node connection (network edge) and minimum confidence of interaction was set at 0.500 (medium-high). Disconnected nodes were discarded. Clustering was performed using k-means and minimum number of clusters of 3. K-parameter (cluster number) was automatically addressed. Functional network enrichment was performed with FDR\u0026nbsp;£\u0026nbsp;1% and strength\u0026nbsp;³\u0026nbsp;0.75. The ranked -log\u003csub\u003e10\u003c/sub\u003e(p-value) significance of change in IGT/NGT protein expression was used as node halo shading in network visualization. Full results of STRING cluster analysis are presented in \u003cstrong\u003eSupplement 2 (Table S9-S11)\u003c/strong\u003e.\u003c/p\u003e"},{"header":"3. Results ","content":"\u003cp\u003e\u003cstrong\u003e3.1 Utilization of\u0026nbsp;\u003c/strong\u003e\u003cstrong\u003em\u003c/strong\u003e\u003cstrong\u003ePAC column for LC/MS/HRMS proteomic analysis of skeletal muscle biopsy.\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe use of 50cm semiconductor-technology micro pillar array columns (mPAC) dictated separate gradient optimization steps, as the direct employment of gradients optimized for grain-based 50 microcapillary columns gave sub-par identification values. Best results were obtained with the use of multi-step (Table S2), 300nl/min 120min gradient (Table S3) at column load of 750ng of skeletal muscle digest (Table S4). Median CV% of retention time variation for all sample-spiked iRT peptides equaled to 26s, with FWHM of 17s (Figure S1 A and B). Pillar array column delivered good chromatographic resolution, retention time stability and reproducible detector signal (Figure S1C and D) accompanied by exceptionally low operating pressure as compared to typical, high-performance, particle-based capillary chromatography column of comparable length (Figure S1D).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e3.2 Initial comparison of methods for ion-chromatogram library generation.\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eAt the early stage of ion-chromatogram library construction, we compared total number of identifications at peptide and protein level for mixed, study-wide muscle sample. HpH fractionation with fraction concatenation (Supplement 1, Figure S6) yielded highest identification numbers, at the level of both the individual and combined fractions. HpH outperformed in this regard both GPF (Supplement 1, Figure S5) and DirectDIA\u003csup\u003eTM\u003c/sup\u003e (Supplement 1, Figure S7) performed on 6 constitutive samples. As the results from HpH and GPF fractionation include runs from 6 independent fractions, we compared them with the values from 6 constitutive non-fractionated sample runs analyzed with DirectDIA\u003csup\u003eTM\u003c/sup\u003e approach. Moreover, to observe the impact of the number of analyzed samples on total number of identifications reported by DirectDIA\u003csup\u003eTM\u003c/sup\u003e we performed 12 constitutive injections, which yielded results comparable to GPF approach (Supplement 1, Figure S5 and S7A).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e3.2 Different approaches to ion-chromatogram library construction for the proteomic analysis of skeletal muscle biopsy.\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eSubsequently, we assembled a number of ion-chromatogram libraries employing only representative fractionation and analysis type techniques or in combination with additional non-fractionated study-wide mixed sample runs (Supplement 1, Table S5), performed with DDA or DIA method. The inclusion of those non-fractionated study-wide mixed runs yielded fractionated/non-fractionated hybrid libraries, with improved retention time alignment across the fractions. Addition of those samples significantly decreased the width of extracted ion-chromatogram (XIC) windows within libraries and improved overall identification number at PSM, peptide and protein level (Supplement 1, Table S5). At this step, we selected HpH-fractionated, DDA-acquired hybrid library, supplemented with 3 non-fractionated DDA samples (HpH/DDA-H library); GPF-fractionated, STW-acquired hybrid library, supplemented with 3 non-fractionated DDA-acquired samples (GPF/STW-H library) and DirectDIA\u003csup\u003eTM\u003c/sup\u003e approach for subsequent analysis (Figure 1, Supplement 1, Table S5).\u003c/p\u003e\n\u003cp\u003eTo further evaluate usefulness of selected approaches in the analysis of skeletal muscle proteome, we compared libraries at the level of PSMs, peptides and proteins. HpH/DDA-H hybrid library excelled in the number of unique PSMs (13771 out of 34740 total), which translated into highest number of unique peptides (8997 out of 23372 total) and proteins (952 out of 2515 total) identified within this library (Figure 2A to C, Supplement 1 Stable S5). Second-best GPF/STW-H library included 44 unique proteins (out of 1573 total), whereas DirectDIA\u003csup\u003eTM\u003c/sup\u003e only 39 (out of 1138 total). Both libraries were almost fully contained within HpH/DDA-H library at the level of peptides and proteins (Figure 2A to C). HpH/DDA-H library included higher number of low-intensity PSMs (Figure 2D), which possibly reflects peptide dilution and loss during fractionation procedure. GPF/STW-H library and DirectDIA\u003csup\u003eTM\u003c/sup\u003e approach yielded higher number of high-intensity PSMs, which in case of the latter could arise from prioritization of high-signal PSMs in de-novo identification pipeline and higher cut-off for low-intensity ones to decrease false-positives. \u0026nbsp;At peptide physicochemical properties level, DirectDIA\u003csup\u003eTM\u003c/sup\u003e approach yielded slightly longer, more hydrophobic peptides, as reflected by their higher mean molecular weight, GRAVY score and hydrophobicity index as compared to both fractionation-based approaches. HpH/DDA-H library based on slightly shorter, hydrophilic peptides. This peptide characteristic could be result of chemical high-pH fractionation and retention of longer, hydrophobic peptides on ethylene bridged hybrid particles of C18 BEH column in high-pH conditions, absorption to plastic during sample transfers, and incomplete solubilization after vacuum concentration characteristic to HpH fractionation pipeline. GPF/STW-H yielded peptides with physicochemical characteristic in between those observed for DirecDIA\u003csup\u003eTM\u003c/sup\u003e and HpH/DDA-H regarding both the peptide length and hydrophobicity (Figure 2E).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e3.4 Evaluation of the ion-chromatogram libraries and DirectDIA\u003csup\u003eTM\u003c/sup\u003e approach for the analysis of diabetes and exercise-related changes in skeletal muscle.\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eTo further evaluate usefulness of respective approaches for the analysis skeletal muscle proteome we performed pathway enrichment analysis using IPA software on proteins identified within a given library, with special emphasis on pathways connected to diabetes, muscle function and energy metabolism. IPA was able to identify 618 significantly enriched pathways in HpH/DDA-H library, covering 37% of their molecular members (pathway fraction) with mean value \u0026nbsp;of 27 proteins per pathway (Figure 2F insert). While GPF/STW-H library presented similar values to HpH/DDA-H ones, DirectDIA\u003csup\u003eTM\u003c/sup\u003e approach yielded considerably lower values for the number of enriched pathways, mean enrichment\u0026nbsp;-log\u003csub\u003e10\u003c/sub\u003e(p-value), pathway fraction and mean proteins per pathway (Figure 2F insert). Detailed enrichment analysis of insulin resistance, protein, energy metabolism, aging and muscular physiology-related pathways reflected above metrics, with HpH/DDA-H library containing highest number of proteins involved – among others - in mitochondrial metabolism, \u0026nbsp;MAPK/AMPK and PI3K/AKT signaling, protein ubiquitination, autophagy, EIF2/eIF4 -controlled protein synthesis and NRF2-mediated oxidative stress response (Figure 2F). GPF/STW-H library yielded second-best results for mitochondrial metabolism, eIF4/p70S6K signaling and protein synthesis/degradation-related pathways, whereas DirecDIA\u003csup\u003eTM\u003c/sup\u003e approach presented lowest number of proteins and enrichment significance in all of the analyzed pathways (Figure 2F).\u003c/p\u003e\n\u003cp\u003eWe concluded library description stage of our study with the replicate (n=3) analysis of mixed, study-wide skeletal muscle sample to observe the impact of respective libraries on PSM, peptide and protein identification metrics. Compared to other non-hybrid \u0026nbsp;library combinations, inclusion of mixed samples decreased by approx. 50% chromatographic XIC window width used for localization of MS/MS peptide fragments (from 5.1 min. to 2.9 min in case of GPF/STW-H library), which translated into higher identification scores noted for all hybrid libraries (Supplement 1, Table S6). Mean number of peptides per protein identified in mixed muscle samples ranged from 8.1 (DirectDIA\u003csup\u003eTM\u003c/sup\u003e) to 9 (HpH/DDA-H). Approx. 94% of proteins present in GPF/STW-H library were identified in mixed muscle samples (Library recovery, Table S6), and each mixed sample presented 92% of total number of proteins identified in triplicate (Completeness, Table S6). Peptide MS/MS spectra from GPF/STW-H library were able to explain approx. 61% of total ion chromatogram signal (Explained TIC, Table S6). For HpH/DDA-H approach, the values for library recovery, sample completeness and explained TIC equaled to 62%, 90% and 65%, respectively, whereas DirectDIA\u003csup\u003eTM\u003c/sup\u003e presented 100%, 100% and 61% for the above parameters. \u0026nbsp;I this case, highest performance at sample completeness and library recovery can be attributed to the nature of this approach, which creates internal ion-chromatogram library from all the samples included in the experiment. This ensures the recovery of complete set of IDs from each sample, albeit with overall lower number of IDs compared to deeper, fractionation-based libraries.\u003c/p\u003e\n\u003cp\u003eSurprisingly, although HpH/DDA-H outperformed other approaches as for muscular proteome coverage at library assembly stage, analysis of skeletal muscle samples with GPF/STW-H library yielded highest unique PSM, peptide and protein identifications, as compared to both the HpH/DDA-H library and DirectDIA\u003csup\u003eTM\u003c/sup\u003e approach (Figure 3A to C, Table S6). Physicochemical characteristics of peptides identified within skeletal muscle samples was similar to the one observed for whole libraries, with HpH and DirectDIA\u003csup\u003eTM\u003c/sup\u003e displaying identification bias towards shorter, hydrophilic and longer, hydrophobic peptides, respectively (Figure 3D). To estimate reproducibility of respective approaches, we analyzed CV% distribution at PSM, peptide and protein level. Surprisingly, most direct DirectDIA\u003csup\u003eTM\u003c/sup\u003e approach without fractionation and library building was characterized by highest CV% at each level of the assay, as compared to both HpH/DDA-H and GPF/STW-H (Figure 3E). At final, protein level, median CV% equaled 15.1% for DirectDIA\u003csup\u003eTM\u003c/sup\u003e, 11% for GPF/STW-H, and 11% for HpH/DDA-H. Ultimately, GPF/STW-H approach yielded highest precision of the assay, with 70% of all proteins measured with CV% below 20%, compared with 69% for HpH/DDA-H and 69 for DirectDIA\u003csup\u003eTM\u0026nbsp;\u003c/sup\u003e(Figure 3E). Protein rank distribution analysis of respective approaches revealed that GPF/STW-H library was able to identify and quantify higher number of proteins at all abundance levels in skeletal muscle samples as compared to other approaches (Figure 3F), with HpH/DDA-H library yielding similar, yet inferior results. Analysis performed with fractionation-based libraries gave significantly (numerically) better results regarding protein rank distribution, which was especially visible for low-abundance proteins (Figure 3F). Quantitative measurements performed with respective approach at displayed good, significant reciprocal correlation (Pearsons r\u0026gt;0.9 with p\u0026lt;0.00001 in all cases) at both the PSM, peptide and protein level. At final, protein level best correlation was observed for GPF/STW-H vs DirectDIA\u003csup\u003eTM\u003c/sup\u003e data (Pearsons r=0.9331, p\u0026lt;0.00001), which could reflect their more direct nature of measurement, compared to chemical fractionation-based HpH/DDA-H approach of library construction. To estimate the feasibility of the particular library in the detection of the muscular proteome alterations induced by glucose intolerance, we performed pathway enrichment analysis using proteins identified by a particular approach. It revealed, that samples analyzed with GPF/STW-H library yielded highest number of significantly enriched pathways (Supplement 1, Figure S8 - table insert), although pathways identified by second-best HpH/DDA-H approach displayed slightly better results regarding the mean significance of enrichment (7.8 vs 7.7\u0026nbsp;-log\u003csub\u003e10\u003c/sub\u003e(p-value)) and mean number of members per enriched pathway (21.4 vs 20.1). Analysis of enrichment of insulin resistance-relevant pathways revealed, that pathways from both the fractionation-based approaches had similar enrichment score (as measured by\u0026nbsp;-log\u003csub\u003e10\u003c/sub\u003e(p-value)) and were equally populated by their molecular members (as measured by the number of identified proteins; Supplement 1, Figure S8). Pathways involved in autophagy, NRF2-mediated oxidative stress response and calcium signaling \u0026nbsp;presented highest enrichment and protein count scores in HpH/DDA-H approach, whereas EIF2-mediated translation control and sirtuin signaling were most pronounced in GPF/STW-H (\u003cstrong\u003eSupplement 1, Figure S8).\u003c/strong\u003e \u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e3.5 Application of fractionation-based library approach or DirectDIA\u003csup\u003eTM\u003c/sup\u003e in the analysis of skeletal muscle proteomic alterations in post-training skeletal muscle from prediabetic subjects.\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e3.5.1 Protein identification and quantitation metrics.\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eFinally, we utilized each of the selected approaches to identify proteomic alterations invoked by glucose intolerance and insulin resistance in skeletal muscle samples from subjects which underwent structured mixed-mode exercise regimen. When applied to the whole experimental IGT/NGT sample set, both HpH/DDA-H and GPF/STW-H approaches were able to identify identical number of protein ratios (Supplement 1, Table S7, Figure S9A to C). Although DirectDIA\u003csup\u003eTM\u003c/sup\u003e excelled at the number of PSMs, it did not translate into higher number of peptides and proteins than those observed for both fractionation-based approaches HpH/DDA-H and GPF/STW-H (1218 vs 1456, respectively). DirectDIA\u003csup\u003eTM\u003c/sup\u003e displayed highest sample completeness (90% vs 64% and 84% for HpH/DDA-H and GPF/STW-H, respectively), library recovery (100% vs 86% and 99%, respectively), peptides per protein (9.8 vs 9.2 and 9.4, respectively), percentage of explained TIC (71% vs 68% and 57%, respectively) and narrowest XIC windows (3.2 min. vs 4.2 and 3.8, respectively). In most of the above quality-related determinants, GPF/STW-H presented second-best results compared to DirectDIA\u003csup\u003eTM\u003c/sup\u003e (Supplement 1 Table S7), simultaneously displaying similar or better numbers at PSM, peptide and protein fold change ratios than the HpH/DDA-H approach (Supplement 1 Table S7, Figure S9A to C). Comparison of identification results for proteins passing 2 unique peptides cutoff, p-value\u0026lt;0.05 cutoff (equivalent to FDR-corrected Q-value of 0.05) and 50% fold change requirement (equivalent to 0.585£-log\u003csub\u003e2\u003c/sub\u003e(FC)£-0.585), revealed that although GPF/STW-H was able to quantify greatest total number of IGT/NGT protein ratios at 2-peptide and p-value cutoff, HpH/DDA-H approach excelled in the number of unique identifications at each of the subsequent steps. Finally, HpH/DDA-H analysis yielded a total of 140 significantly affected proteins passing all of the cutoffs, compared to 131 for GPF/STW-H \u0026nbsp;and 64 for DirectDIA\u003csup\u003eTM\u003c/sup\u003e (Supplement 1, Figure S9D to F). Protein abundance rank analysis also confirmed higher proficiency of both fractionation-based approaches in protein identification compared to DirectDIA\u003csup\u003eTM\u0026nbsp;\u003c/sup\u003e (Supplement 1 Figure S9G), although the latter one was able to quantify 29 more IGT/NGT ratios at the lowest protein abundance compared to both the HpH/DDA-H and STW/DIA-H.Study-wide correlation analysis of protein abundance between respective approaches yielded high Pearson r values for both the IGT-only and NGT-only samples (r\u0026gt;0.92, p\u0026lt;0.00001 in all cases), with HpH/DDA-H and STW/DIA-H displaying highest correlation of respective protein expression (Supplement 1, Figure S10A). Correlation analysis of IGT/NGT differential protein expression (expressed as protein log2 fold change, log2FC) yielded significant (p\u0026lt;0.00001) correlations \u0026nbsp;between all of the approaches at the respective 2-peptide (Pearson r\u0026gt;0.55, all cases) and Q-value (Pearson r\u0026gt;0.7, all cases) cutoffs \u003cstrong\u003e(Supplement 1, Figure S10B)\u003c/strong\u003e. Introduction of 50% fold-change cutoff drastically reduced the number of proteins shared between appropriate approaches, yet yielded highest Pearson r values of 0.91 and 0.96 for STW/DIA-H vs DirectDIA\u003csup\u003eTM\u003c/sup\u003e and HpH/DDA-H vs DirectDIA\u003csup\u003eTM\u003c/sup\u003e, respectively (p\u0026lt;0.00001). Despite highly correlated results of protein expression between HpH/DDA-H and STW/DIA-H at the level of IGT-only and NGT-only samples, the final log2FC values for significantly affected proteins displayed modest, although significant Pearson correlation of 0.7 \u003cstrong\u003e(Supplement 1, Figure S10B)\u003c/strong\u003e. Importantly, when compared between approaches, the shared proteins passing all 3 cutoffs and those passing 2-peptide and p-value cutoff (with minor exceptions) displayed the same direction and similar degree of differential regulation \u003cstrong\u003e(Supplement 1, Figure S10B)\u003c/strong\u003e.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e3.5.2 Molecular pathways enrichment analysis\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eSubsequently, using significant-only proteins (passing all 3 cutoffs), we performed pathway enrichment analysis and protein functional clustering, to identify proteomic alterations evoked by prediabetic state in post-training muscle and to evaluate the usefulness of each approach in the identification of the above changes. Quantitatively, GPF/STW-H identified highest number of significantly affected pathways (233) compared to both HpH/DDA-H and DirectDIA\u003csup\u003eTM\u003c/sup\u003e (200 and 98, respectively). DirectDIA\u003csup\u003eTM\u003c/sup\u003e performed better than both fractionation-based approaches, regarding mean pathway enrichment and proteins/pathway metrics (Figure 4, table insert). All of the approaches were equally effective in the detection of alterations in EIF2-mediated protein synthesis in IGT group, yet only DirectDIA\u003csup\u003eTM\u003c/sup\u003e was able to identify significant changes in protein ubiquitination, ubiquitin-like FAT10 signaling and detected changes in autophagy pathway with higher sensitivity than other approaches \u003cstrong\u003e(Figure 4A, Supplement 2 Table 6 to 8)\u003c/strong\u003e. Higher sensitivity of DirectDIA\u003csup\u003eTM\u003c/sup\u003e was also noted for all of the studied Aging/Stress/ROS-related pathways, such as sirtuin pathway, BAG2-mediated stress response and NRF2 oxidative stress response. Compared to both fractionation-based techniques, DirectDIA\u003csup\u003eTM\u003c/sup\u003e was unable to identify changes in pathways related to skeletal muscle physiology and contractile function, in which GPF/STW-H approach displayed best performance \u003cstrong\u003e(Figure 4A)\u003c/strong\u003e. Interestingly, all techniques were unable to detect in IGT group alterations in molecular pathways commonly connected with insulin resistance and prediabetes, such as mitochondrial dysfunction, PKA and AMPK and Type II DM-related signaling, which suggest that above hallmarks of diabetic state were normalized in the muscle of trained prediabetics. Significant changes were observed in protein synthesis-related eIF4 and p70S6K pathways, glucocorticoid signaling, and ERK/MAPK signaling, yet only GPF/DIA-H approach was able to detect alterations in PI3K/AKT-controlled pathway.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eTo detect directionality of the changes, we performed z-score analysis on the significantly affected proteins, which revealed that IGT post-training muscle displayed down-regulation (z-score\u0026lt;2) in a total of 7 molecular pathways responsible for control of the protein synthesis, compared to post-training NGT counterparts (selenoaminoacid metabolism, detection of amino-acid deficiency, translation initiation, elongation, termination, RNA processing and EIF2 signaling) \u003cstrong\u003e(Figure 4B, Supplement 2 Table 6 to 8)\u003c/strong\u003e. Both GPF/STW-H and DirectDIA\u003csup\u003eTM\u003c/sup\u003e were equally sensitive in the detection of the above alterations (with EIF2 pathway as an exception), whereas HPH/DDA-H approach identified only 3 out of 7 in total. Contrary to DirectDIA\u003csup\u003eTM\u003c/sup\u003e, both fractionation-based techniques detected significant down-regulation in the expression of proteins involved in skeletal muscle contraction (at the level of actin cytoskeleton, integrin signaling and muscle hypertrophy response, Figure 4B), whereas only DirectDIA\u003csup\u003eTM\u003c/sup\u003e was able to detect detrimental alterations in pathways responsible for DNA synthesis, replication and repair and PTEN-dependent regulation of the cell cycle \u003cstrong\u003e(Figure 4B, Supplement 2 Table 6 to 8)\u003c/strong\u003e. Only GPF/STW-H approach was sensitive enough to detect down-regulation in the PI3K/AKT, a key insulin signaling pathway.\u003c/p\u003e\n\u003cp\u003eRegarding structural and functional alterations which distinguish post-training IGT group from their normoglycemic counterparts, both HpH/DDA-H and GPF/DIA-H methods detected greater number of changes compared to DirectDIA\u003csup\u003eTM\u003c/sup\u003e \u003cstrong\u003e(Figure 4C, Supplement 2 Table 6 to 8)\u003c/strong\u003e. Interestingly, whereas HpH/DDA-H was able to detect significant changes in lipid metabolism-related muscle functions (lean body mass retention, quantity of lipid droplets) and myofiber differentiation, GPF/DIA-H excelled in detection of all muscle contractile and morphology-related changes, such as altered Ca\u003csup\u003e2+\u003c/sup\u003e storage, abnormal morphology of muscle and its fibers, strength of contraction and disrupted motor plate function \u003cstrong\u003e(Figure 4C, Supplement 2 Table 6 to 8)\u003c/strong\u003e. DirectDIA\u003csup\u003eTM\u003c/sup\u003e detected molecular alterations connected with density of neuromuscular junctions and skeletal muscle fibrosis.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e3.5.2 Protein functional clustering\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eTo further investigate the differences between IGT and NGT subjects observed in post-training skeletal muscle, and to identify relationships between differentially regulated proteins, we performed functional protein clustering analysis with each of the different approaches. Total of 103 out of 141 significantly affected proteins identified by HpH/DDA-H method (74% of total) generated 4 major interconnected clusters, all implicated in different aspects of protein metabolism \u003cstrong\u003e(Figure 5, Supplement 2, Table S9)\u003c/strong\u003e. Cluster I encompassed proteins involved in the regulation of DNA expression (MAPK kinase group), mRNA processing (HNRNP proteins), translation control (EIF factors of initiation, elongation and termination) and ribosomal subunits. Cluster II aggregated proteins involved in ER-mediated protein processing and post-translational modifications, whereas Cluster III and VI gathered proteins involved in Golgi-mediated vesicular transport and endocytic vesicular recycling. Those findings indicate that HPH/DDA-H was able to detect significant alterations in molecular control of protein synthesis, processing and trafficking in trained pre-diabetics compared to respective normoglycemic counterparts. This observation explains the presence of \u0026nbsp;Cluster IV, composed of muscle contractile apparatus and extracellular matrix proteins, which suggest abnormal skeletal muscle composition in post-training IGT group compared to NGT one. Finally, HPH/DDA-H approach identified a known hallmark of prediabetes i.e. alterations in proteins connected with mitochondrial fatty acids metabolism (Cluster V).\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eSimilarly, clustering of 85 out of 130 (65%) of significantly affected proteins revealed by GPF/DIA-H emphasized alterations in mRNA processing, translation and protein metabolism (Cluster II and its 2 subclusters, which included MAPK kinases, EIF translation control factors, ribosomal subunits, RAS pathway members and aminopeptidases) and muscle contraction and extracellular matrix (Cluster III, which included light and heavy myosin chain isoforms, actin and EC matrix laminin B1, decorin and prolagin) \u0026nbsp;\u003cstrong\u003e(Figure 6, Supplement 2, Table S10)\u003c/strong\u003e. Interestingly, unique finding of GFP/DIA-H approach was the identification of clusters, which molecular members are involved in branched amino-acids biosynthesis and degradation (Cluster IV e.g. mitochondrial BCAT2 branched chain aminotransferase, \u0026nbsp;ACADBS short/branched chain acyl-CoA dehydrogenase, MCEE mitochondrial methylmalonyl-CoA epimerase), skeletal muscle carnosine metabolism (Cluster IV, e.g. ) and pyridoxal phosphate (vitamin B6) metabolism (Cluster V, i.e. PDXK pyridoxal kinase and PLPBP pyridoxal phosphate binding protein). Alterations in BCAA metabolism, muscular carnosine dipeptide content and vitamin B6 deficiency display strong correlation with prediabetes and subsequent progression toward Type 2 DM. Clustering of the most diverse protein set was noted for Cluster I \u003cstrong\u003e(Figure 6, Supplement 2, Table S10)\u003c/strong\u003e, and included proteins connected with gene expression and mRNA processing at mitochondrial and nuclear level (e.g. TFAM mitochondrial transcription factor A, proteins from HRNR heterogeneous nuclear ribonucleoprotein family, SNRBP RNA spliceosome protein and STRAP ribonucleoprotein assembly protein), cytoskeleton/plasma membrane interaction and assembly proteins (e.g. caveolae associated CAV1, CAVIN4, STIM1 proteins and cytoskeleton assembly and plasma membrane anchoring FLNA flaminin, MSN moesin, SNTB1\u0026nbsp;b-1-syntropin and ANK3 Ankyrin-3 proteins) lipid transport and metabolism-associated proteins (APOE, APOC1 lipoproteins and FABP4 fatty acids binding protein) and finally ROS metabolism and pentose phosphate pathway proteins (GPX3 glutathione peroxidase 3, TXN2 mitochondrial thioredoxin and H6PD hexose-6-phosphate dehydrogenase/glucose 1-dehydrogenase, PGLS 6-phosphogluconolactonase, respectively). APOE played the central, linking role in formation of Cluster I, possibly due to its impact on caveolae function, lipid metabolism, and mitochondrial dysfunction function \u003cstrong\u003e(Figure 6, Supplement 2, Table S10).\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eProtein metabolism-related clustering of significantly affected proteins was also noted for DirectDIA\u003csup\u003eTM\u003c/sup\u003e analyzed IGT vs NGT dataset. A total of 38 out of 64 proteins (65%) yielded 6 independent (not connected) clusters, with Cluster I, II and IV aggregating proteins involved in muscle contraction, cytoskeleton formation and extracellular matrix (Cluster I, e.g mysin and actin isoforms, collagens, serpin, laminin) mRNA processing, translation and protein degradation (Cluster II e.g) and protein folding and glycosylation (Cluster V e.g. calumenin, DDOST and RPN2 glycosyltransferases). Interestingly, Cluster II included both the molecular members of protein synthesis pathways (EIF factors, ribosomal subunits) and proteasomal degradation pathway (PSM proteasomal proteins) (\u003cstrong\u003eFigure 7, Supplement 2, Table S11)\u003c/strong\u003e, which was unique finding of DirectDIA\u003csup\u003eTM\u003c/sup\u003e analysis. Cluster III connected proteins involved in muscular calcium binding and Ca\u003csup\u003e2+\u003c/sup\u003e regulation of muscle contraction (S100 family proteins, parvalbumin) whereas 2 members of Cluster V shared the same molecular function as \u0026nbsp;acyl-CoA – synthesizing enzymes (mitochondrial medium and short-chain Acyl-CoA ligases). \u0026nbsp;\u003c/p\u003e"},{"header":"4. Discussion","content":"\u003cp\u003eIntroduction of high-throughput proteomic analysis in clinical applications carries a promise of possible ground-breaking discoveries at the level of both the foundational science and medical applications. Initial attempts based on 2DGE with MALDI-TOF \u0026nbsp;peptide identification or FT-ICR LC/MS ultra-high resolution mass spectrometry, although encouraging from the basic science standpoint, were lacking in important qualities, such as sample throughput, reproducibility and proteome coverage. Additionally, plasma \u0026ndash; a readily-available clinical sample - presents significant analytical challenge regarding protein matrix composition (high-abundance protein \u0026ldquo;iceberg\u0026rdquo; effect). Those shortcomings were the result of several bottlenecks and variabilities at almost every step of the proteomic pipeline, including cumbersome sample preparation, consistency of in-house prepared nanoLC columns, slow-cycle DDA-based acquisition and early software employed in peptide identification and protein quantitation. Current generation of MS-based proteomic pipelines finally resolve all the above issues. Moreover, partial application of modern pipelines on previous-generation mass spectrometers allows for the high-throughput proteomic analysis of quality and reproducibility required for clinical applications. Taking in to account all of the above, we employed modern DIA-based approach consisting of staggered-window MS/MS acquisition, customized ion-chromatogram libraries and highly reproducible semiconductor-technology based micro pillar array columns coupled with Orbitrap-HRMS for the analysis of challenging skeletal muscle samples. Fine-tuning of\u0026nbsp;mPAC chromatography, DIA acquisition and library construction allowed for elucidation of skeletal muscle proteome changes in prediabetic patients, confirming presence of persistent, adverse changes in muscle proteome, despite 3 months of structured exercise.\u003c/p\u003e\n\u003cp\u003eCompared with basic science proteomics research, which can afford for multiple re-analysis attempts and longer analysis times to achieve best possible data quality, proteomics in clinical applications puts emphasis on robustness, batch-to-batch reproducibility and throughput of the complete analysis pipeline. To improve robustness and reproducibility of LC separations we employed micro pillar array columns which \u0026ndash; compared to particle based counterparts \u0026ndash; display decreased batch-to batch variability, significantly lower backpressures, full flow reversibility and low carryover \u003csup\u003e43\u003c/sup\u003e. Those features arise from semiconductor-type manufacturing process, which creates 5mm ID pillars, 18\u0026mu;m in height separated by 2.5\u0026mu;m gaps \u003csup\u003e27\u003c/sup\u003e. In our study, 1\u003csup\u003est\u003c/sup\u003e Gen 50cm\u0026nbsp;mPAC column paired with matching 1cm\u0026nbsp;mPAC trap generated 20x lower backpressure than particle-based counterpart (25bar vs 500bar) and negligible pressure drop during trap valve operation, increasing robustness of HPLC analysis. The column displayed good retention time stability, and was resistant to clogging even in case of samples rich in fibrous biological polymers (skeletal muscle). In case of increased backpressure,\u0026nbsp;mPAC column could be reverse-flushed to restore its original performance, increasing column longevity. This cannot be applied to packed, particle-based nanoflow capillary columns without the risk of stationary phase loss. Currently, pillar array columns are increasingly used in proteomics applications\u0026nbsp;\u003csup\u003e28,44-47\u003c/sup\u003e with 2\u003csup\u003end\u003c/sup\u003e generation columns (2.5mm ID pillars, 25mm height 1.5\u0026mu;m interpillar distance) displaying improved characteristic regarding theoretical plate heigh and resolution\u0026nbsp;\u003csup\u003e27,28\u003c/sup\u003e. Clinical applications can benefit from newest, rectangle pillar-based design generation (75mm x 3mm pillars), which despite its short 5.5cm length allows for higher-throughput,\u0026nbsp;mLC-like analysis, with separation power comparable to its 50cm circular pillar array counterpart\u0026nbsp;\u003csup\u003e44,48\u003c/sup\u003e.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eLow column-to-column variability and retention time consistency displayed by micro pillar array column are crucial in the robust implementation of DIA analysis in both the library-based and pure bioinformatic approaches. For our study we selected staggered windows approach (STW-DIA) for both the gas-phase fractionation library generation and experimental runs, first introduced by Searle, Pino and Amodei \u003csup\u003e37-39\u003c/sup\u003e. Although non-overlapped window placement would yield increased MS/MS cycle times, in our view the window staggering is better suited for older-generation quadrupole-orbitrap instruments, due to the correction of non-rectangular isolation characteristic of quadrupoles and increased precision of MS/MS measurements crucial for DIA-based assays. Despite being rarely employed, this approach is gaining increasing acceptance and was employed in recent studies on both previous generation and modern mass spectrometers \u003csup\u003e30,49\u003c/sup\u003e.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eTo elucidate the impact of ion-chromatogram library assemble on the results of DIA-STW based analysis of exercise-induced changes in prediabetic subjects, we selected 3 approaches with increasing deepness of the library coverage, namely High-pH fractionation (HpH), gas-phase fractionation (GPF) and pure computational approach (Spectronaut\u003csup\u003eTM\u003c/sup\u003e DirectDIA\u003csup\u003eTM\u003c/sup\u003e). Post-digestion high-pH fractionation with fraction concatenation (HpH) was shown multiple times to generate most comprehensive ion-chromatogram libraries, as compared to other protein and peptide fractionation techniques, such as PAGE, IEF focusing variants, HILIC, SCX and SAX ion exchange \u003csup\u003e50-52\u003c/sup\u003e. When performed at microflow ranges (50-100\u0026mu;l/min) with 100\u0026mu;g of total protein digest, we were able to perform all the necessary steps such as fraction, fraction concatenation, vacuum concentration and subsequent LC/MS/MS analysis in a single 96 well-plate format, which assures low sample loss. Nevertheless, HPLC-based physical fractionation is time-consuming, requires additional instrumentation (vacuum concentrators, dedicated HPLC with fraction collection or multimode nanoLC/microLC HPLC system etc.). In-source gas-phase fractionation offers cost and time-effective alternative to physical peptide fractionation, requiring only several additional analysis runs, albeit with lower library coverage. To further increase the deepness proteome analysis, staggered-windows gas-phase fractionation was recently combined by Penny et al. with ion-mobility technique (TIMS-diaPASEF), yielding significant improvements over its basic versions \u003csup\u003e53\u003c/sup\u003e. Finally, pure bioinformatic solutions to DIA analysis rely on direct identification of peptide product ions without the prior assembly of ion-chromatogram libraries. Currently, this approach to DIA analysis displays fastest growth in capabilities and multitude of available software solutions, presenting different analytical approaches (e.g spectrum-centric, peptide-centric, in-silico fragmentation libraries etc.). Recently, pure computational DIA analysis was updated with machine learning, neural networks and AI capabilities (for excellent reviews and comparative studies see \u003csup\u003e30,54\u003c/sup\u003e and \u003csup\u003e33,55\u003c/sup\u003e, respectively). Direct analysis of DIA-acquired MS/MS runs presents the fastest, direct approach, yet requires significant computational resources. Moreover, contrary to both the fractionation-based libraries which, ideally, should be prepared with the use of study-wide mixed sample, it can be applied continuously, in parallel to sample collection and processing before the completion of the study.\u003c/p\u003e\n\u003cp\u003eOur primary objective was to identify the differences of 3 major approaches on the outcome of DIA-based analysis of muscular proteome changes evoked by exercise in prediabetic patients. We hypothesized, that assembly of spectral libraries can have direct impact on the outcome of the analysis and the final biological findings. Firstly, we selected best performing libraries among several possible combinations. Interestingly, hybrid libraries which in addition to GPF or HpH fraction runs included full mass range DDA runs from study-wide mixed samples displayed narrower XIC search windows, which translated into lower CV% of replicate analysis. Greatest improvement was noted for both of the fractionated libraries, which in our opinion is connected with the improved RT alignment of across particular library fractions, as only non-fractionated samples contained all of the proteins and peptides. Surprisingly, inclusion of non-fractionated study-wide sample had greater impact on RT alignment and XIC window width, than the presence of iRT peptides in each of the fractions, which in theory should provide RT anchor points across library runs. Although this improvement should be most visible in the case of GPF hybrid library (as iRT peptides fall in different mass fractions), it was also noted for HpH hybrid library, where iRT peptides were added to concatenated fractions prior to LC/MS runs. As expected, DirectDIA\u003csup\u003eTM\u003c/sup\u003e approach displayed the best RT alignment and narrowest XIC windows, yet yielded greatest CV% of replicate analysis. This outcome was somehow surprising, as protein rank distribution and precursor intensity distribution suggested, that DirectDIA\u003csup\u003eTM\u003c/sup\u003e displayed bias toward higher intensity signals, inherently easier to quantify. Moreover, DirectDIA\u003csup\u003eTM\u003c/sup\u003e analysis targeted longer, more hydrophobic peptides, compared to analysis performed with HPH/DDA-H library, which could be explained by hydrophobic peptide loss during HpH fractionation. Yet the difference was also visible between DirectDIA\u003csup\u003eTM\u003c/sup\u003e and \u0026nbsp; GPF/STW-H analysis, which cannot be explained by mass-range gas-phase fractionation. We observed this phenomenon in both the results of the analysis of mixed-sample replicates and whole sample. Those findings suggest, that biological outcomes of the analysis performed with different approaches could differ due to targeting of peptides with different physicochemical properties, displaying slight bias toward membrane or soluble proteins.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eAs expected, HpH/DDA-H ion-chromatogram library contained the highest number of identified proteins and presented most populated pathways relevant to T2D and insulin resistance. \u0026nbsp;Yet this advantage did not translate into higher-quality results, when employed to quantify proteins in study-wide mixed sample. The GPF/STW-H \u0026ndash; based results were similar or better regarding quantity of proteins and their number included in T2D-relevand pathways. This gives significant advantage for staggered-windows gas-phase fractionation approach, which requires less time and resources for implementation.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eRegarding the outcome of the analysis performed on whole set of experimental samples, we noted both similarities and significant differences depending on the use of particular approach. Although we observed strong, group-wise correlations between individual proteins in IGT or NGT groups quantified with 3 different approaches (Figure S10A), the final IGT vs NGT differential expression ratios displayed weaker association (Figure S10B). Moreover, each approach turned up relatively different set of proteins, when all significance cut-offs were considered (\u0026gt;= 2 peptides, Q-value\u0026lt;0.05, 50% FC), with only 8 common proteins (down-regulated in IGT group: TAGLN, MYBPH, KRT2, S100A13, PRELP ACADSB and up-regulated in IGT group C4A and ACTN3) common between all 3 approaches. Those discrepancies arise from differences in FDR-corrected p-values (Q-values) and calculated differential expression ratios, as those parameters substantially decreased the number of shared proteins. Interestingly, both the molecular function and the direction of regulation of some of those proteins align with the metabolic and functional deficiencies observed in insulin-resistant muscle. Down-regulation of MYBPH, which is highly expressed in insulin-sensitive \u003csup\u003e56,57\u003c/sup\u003e, mitochondria-rich type 1 oxidative muscle fibers \u003csup\u003e58\u003c/sup\u003e, can reflect the decrease of fiber type in IGT group muscle, despite structured exercise regimen. Similarly, down-regulation of S100A13 calcium-binding protein, which over-expression is connected with mitochondrial biogenesis in hypoxia-trained skeletal muscle \u003csup\u003e59\u003c/sup\u003e, aligns with decreased oxidative capacity and mitochondrial content observed in insulin resistant muscle \u003csup\u003e60\u003c/sup\u003e, whereas decrease in expression of mitochondrial ACADSB branched-chain dehydrogenase can be traced to both mitochondrial deficiency \u003csup\u003e61\u003c/sup\u003e and disrupted branched chain AA metabolism in prediabetic subjects\u003cstrong\u003e\u0026nbsp;\u003c/strong\u003e\u003csup\u003e62,63\u003c/sup\u003e. Regarding over-expressed proteins, complement protein C4A up-regulation corresponds with increased pro-inflammatory response observed in obesity-induced insulin resistance \u003csup\u003e64,65\u003c/sup\u003e. Taking into account all significantly affected proteins, each approach was more sensitive towards particular changes observed in skeletal muscle proteome of IGT group. DirectDIA\u003csup\u003eTM\u003c/sup\u003e analysis identified pathways involved in proteasomal protein degradation, cell cycle control and DNA replication and repair, which was not observed in both fractionation-based approaches. Moreover, DirectDIA\u003csup\u003eTM\u003c/sup\u003e was more sensitive towards identification of changes connected with protein synthesis and degradation, oxidative stress and sirtuin signaling pathways, although all approaches signaled decreased expression of proteins involved in translation control. GPF/STW-H library based analysis was more sensitive towards detecting changes in skeletal muscle calcium signaling and morphology, whereas HpH/DDA-H in eIF4 and p70S6 kinase signaling and decreased muscle hypertrophy. Differences in biological data interpretation were also visible at the level of protein functional clustering, with DirectDIA\u003csup\u003eTM\u003c/sup\u003e detecting unique alternations in proteins involved in PTM glycation, Ca\u003csup\u003e2+\u003c/sup\u003e binding and acyl-CoA synthesis, GPF/STW-H in proteins involved in branched amino-acid metabolism and vitamin B6 metabolism, whereas HpH/DDA-H in ER protein processing \u0026nbsp;and Golgi vesicular transport. All of the approach-dependent unique protein clusters are important for metabolic function of skeletal muscle in normoglycemia, as alternations in muscular Ca\u003csup\u003e2+\u003c/sup\u003e metabolism \u003csup\u003e66,67\u003c/sup\u003e\u0026nbsp; and protein glycation \u003csup\u003e68\u003c/sup\u003e, branched-chain AA \u003csup\u003e62,63\u003c/sup\u003e and vitamin B6 metabolism \u003csup\u003e69,70\u003c/sup\u003e and proteins-synthesis related ER stress \u003csup\u003e71,72\u003c/sup\u003e are hallmarks of insulin resistance and T2DM. Nevertheless, significant alternation observed in trained IGT group as compared to respective NGT group, was the down-regulation of pathways involved in protein synthesis, despite 3 months of controlled exercise regimen. Adverse effects of insulin resistance on skeletal muscle protein synthesis could be responsible for the observed decrease in sarcomere and extracellular matrix proteins, leading to skeletal muscle contractile dysfunction \u003csup\u003e17,18\u003c/sup\u003e. Regarding both the qualitative and quantitative aspects of particular approach, HpH/DDA-H \u0026ndash; based analysis identified the greatest number of significant proteins and generated unique protein interaction networks, covering important aspects of skeletal muscle metabolism such as ER stress and Golgi protein processing. On the other hand GPF/STW-H approach presented similar quantitative performance, identified significant portion of affected proteins and unique features of muscular insulin resistance (disturbance in branched AA metabolism), being less problematic to implement. Finally, libraryless DirectDIA\u003csup\u003eTM\u003c/sup\u003e analysis, although presented fewer significantly affected proteins, was able to robustly detect common features of skeletal muscle insulin resistance, e.g. disturbed control of protein synthesis and proteasomal protein degradation. Taking all of the above, \u0026nbsp;gas-phase fractionated library, acquired in staggered windows mode and supplemented with DDA full mass range runs constitute an attractive alternative for time- and resource-consuming physical peptide fractionation. Moreover, identification and quantitation of whole sample set using GPF/STW-H approach was computationally less demanding than pure computational approach, moreover doubling the number of significant proteins. Although improvements in pure computational identification and quantitation will narrow the gap between library-based and libraryless approaches, the similar augmentation of library-based quantitation will further advance DIA-based assay.\u003c/p\u003e"},{"header":"5. Conclusions","content":"\u003cp\u003eAmong multitude of MS acquisition modes employed in proteomic assays, data-independent analysis coupled with robust chromatographic separation is best suited for large projects performed on clinical samples. Important aspect of DIA-based assay is the assembly of ion-chromatogram libraries or selection of pure computational strategy for DIA data interrogation, which hypothetically can lead distinct biological interpretations. Our goal was to employ different strategies to library construction to observe their impact on the DIA-based analysis of post-exercise differences in skeletal muscle proteome between glucose intolerant and healthy subjects. Although all tested approaches were able to detect alternations in control of protein synthesis and sarcomere protein expression, each one identified unique changes important to \u0026nbsp;metabolic and contractile muscle function. In our view, staggered-windows gas-phase fractionated ion-chromatogram library presented the best balance between reproducibility, detection of significant changes, depth of analysis and ease of implementation. Combined together, the biological findings indicate, that despite extensive structured exercise regimen, insulin-resistant muscle displays disturbed molecular pathways implicated in protein synthesis, intracellular trafficking and processing, branched amino-acids and acyl-CoA mitochondrial metabolism and calcium balance, which can be the cause of muscular contractile dysfunction observed in insulin resistant subjects.\u0026nbsp;\u003c/p\u003e"},{"header":"Abbreviations","content":"\u003cp\u003e2DPAGE \u0026ndash; 2 dimensional polyacrylamide gel electrophoresis\u003c/p\u003e\n\u003cp\u003eABC - Ammonium Bicarbonate\u003c/p\u003e\n\u003cp\u003eACG - Automatic Gain Control\u003c/p\u003e\n\u003cp\u003eACN - Acetonitrile\u003c/p\u003e\n\u003cp\u003eAMPK - AMP-Activated Protein Kinase\u003c/p\u003e\n\u003cp\u003eAU - Arbitrary Units\u003c/p\u003e\n\u003cp\u003eAUC - Area Under the Curve\u003c/p\u003e\n\u003cp\u003eBCA - Bicinchoninic Acid\u003c/p\u003e\n\u003cp\u003eBCAA - Branched-Chain Amino Acids\u003c/p\u003e\n\u003cp\u003eBMI - Body Mass Index\u003c/p\u003e\n\u003cp\u003eCa\u0026sup2;⁺ - Calcium Ion\u003c/p\u003e\n\u003cp\u003eChol - Cholesterol\u003c/p\u003e\n\u003cp\u003eDDA - Data-Dependent Acquisition\u003c/p\u003e\n\u003cp\u003eDIA - Data-Independent Acquisition\u003c/p\u003e\n\u003cp\u003eDirectDIA\u0026trade; - Direct (libraryless) Analysis of Spectra form Data-Independent Acquisition\u003c/p\u003e\n\u003cp\u003eDPP4 - Dipeptidyl Peptidase 4\u003c/p\u003e\n\u003cp\u003eEIF2 - Eukaryotic Initiation Factor 2\u003c/p\u003e\n\u003cp\u003eER - Endoplasmic Reticulum\u003c/p\u003e\n\u003cp\u003eFA - Formic Acid\u003c/p\u003e\n\u003cp\u003eFDR - False Discovery Rate\u003c/p\u003e\n\u003cp\u003eFT-ICR LC/MS - \u0026nbsp;Liquid chromatography/Fourier transform ion cyclotron resonance mass spectrometry\u003c/p\u003e\n\u003cp\u003eFPG - Fasting Plasma Glucose\u003c/p\u003e\n\u003cp\u003eGLP1 - Glucagon-Like Peptide 1\u003c/p\u003e\n\u003cp\u003eGO - Gene Ontology\u003c/p\u003e\n\u003cp\u003eGPF - Gas-Phase Fractionation\u003c/p\u003e\n\u003cp\u003eHbA1c - Hemoglobin A1c\u003c/p\u003e\n\u003cp\u003eHDL - High-Density Lipoprotein Cholesterol\u003c/p\u003e\n\u003cp\u003eHOMA2 %B - Homeostatic Model Assessment of Beta Cell Function\u003c/p\u003e\n\u003cp\u003eHOMA2 %S - Homeostatic Model Assessment of Insulin Sensitivity\u003c/p\u003e\n\u003cp\u003eHOMA2-IR - Homeostatic Model Assessment for Insulin Resistance\u003c/p\u003e\n\u003cp\u003eHOMAD - Homeostatic Model Assessment for Diabetes\u003c/p\u003e\n\u003cp\u003eHpH \u0026ndash; Reverse-phase HPLC Peptide Fractionation in high-pH conditions\u003c/p\u003e\n\u003cp\u003eHRMS - High-Resolution Mass Spectrometry\u003c/p\u003e\n\u003cp\u003eIAA - Iodoacetamide\u003c/p\u003e\n\u003cp\u003eIFG - Impaired Fasting Glucose\u003c/p\u003e\n\u003cp\u003eIGT - Impaired Glucose Tolerance\u003c/p\u003e\n\u003cp\u003eINS - Insulin\u003c/p\u003e\n\u003cp\u003eiRT \u0026ndash; Stable Isotope-labelled peptides for Peptide Indexed Retention Time calculation\u003c/p\u003e\n\u003cp\u003eLC/MS/MS - Liquid Chromatography/Tandem Mass Spectrometry\u003c/p\u003e\n\u003cp\u003eLDL - Low-Density Lipoprotein Cholesterol\u003c/p\u003e\n\u003cp\u003eLN₂ - Liquid Nitrogen\u003c/p\u003e\n\u003cp\u003eMALDI-TOF - Matrix-assisted laser desorption/ionization \u0026ndash; time of flight mass spectrometry\u003c/p\u003e\n\u003cp\u003eMSX - Multiplexed MS/MS\u003c/p\u003e\n\u003cp\u003eNCE - Normalized Collision Energy\u003c/p\u003e\n\u003cp\u003eNGT - Normal Glucose Tolerance\u003c/p\u003e\n\u003cp\u003eOGTT - Oral Glucose Tolerance Test\u003c/p\u003e\n\u003cp\u003ePSM - Peptide-Spectrum Match\u003c/p\u003e\n\u003cp\u003ePTM/PTMs - Post-Translational Modification/Post-Translational Modifications\u003c/p\u003e\n\u003cp\u003eSDC - Sodium Deoxycholate\u003c/p\u003e\n\u003cp\u003eSGLT2 - Sodium-Glucose Co-Transporter 2\u003c/p\u003e\n\u003cp\u003eSMM - Skeletal Muscle Mass\u003c/p\u003e\n\u003cp\u003eSTRING - Search Tool for the Retrieval of Interacting Genes/Proteins\u003c/p\u003e\n\u003cp\u003eSTW - Staggered Windows\u003c/p\u003e\n\u003cp\u003eT2DM - Type 2 Diabetes Mellitus\u003c/p\u003e\n\u003cp\u003eTCEP - Tris(2-carboxyethyl)phosphine\u003c/p\u003e\n\u003cp\u003eTFA - Trifluoroacetic Acid\u003c/p\u003e\n\u003cp\u003eTG - Triglycerides\u003c/p\u003e\n\u003cp\u003eTIC - Total Ion Chromatogram\u003c/p\u003e\n\u003cp\u003eVAT_DXA - Visceral Adipose Tissue measured by Dual-Energy X-ray Absorptiometry\u003c/p\u003e\n\u003cp\u003eVO₂max - Maximal Oxygen Uptake\u003c/p\u003e\n\u003cp\u003e\u0026mu;PAC\u0026trade; - Micro Pillar Array Column\u003c/p\u003e"},{"header":"Declarations","content":"\u003ch2\u003eAuthor Contribution\u003c/h2\u003e\u003cp\u003eA.C., L.S. and P.Z designed the study; L.S supervised subjects-related parto of the study and sample collection stage; A.C-R and P.K assisted in sample anthropometric data collection; A.C. and P.Z performed all sample analysis, data analysis, data interpretation and statistical analysis; M.C. assisted in data analysis nad statistical analysis; A.C and P.Z. wrote original draft manuscript and prepared all manuscript figures; A.B-Z . supplied resources, performed all diabetes-related data interpretation; P.Z., L.S. and A.K. supplied funding. All authors reviewed the manuscript.\u003c/p\u003e\u003cp\u003e\u003cstrong\u003eData availability:\u003c/strong\u003e The mass spectrometry proteomics data have been deposited to the ProteomeXchange Consortium (http://proteomecentral.proteomexchange.org) via the PRIDE partner repository \u003csup\u003e73\u003c/sup\u003e with the dataset identifier PXD055536.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eReviewer account details:\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eProject accession:\u0026nbsp;\u003c/strong\u003ePXD055536\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eUsername:\u0026nbsp;\u003c/strong\[email protected]\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003ePassword:\u0026nbsp;\u003c/strong\u003epIpjyglKQ2WB\u003c/p\u003e\n\u003cp\u003eThis study was supported by funds from the Ministry of Education and Science of Poland, within the project “Excellence Initiative—Research University”, Ministry of Health of Poland within the project “Center for Artificial Medicine at the Medical University of Bialystok” and the Medical University of Bialystok grant B.SUB.24.393\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\u003cli\u003e\u003cspan\u003eD, V. et al. Projecting the economic burden of type 1 and type 2 diabetes mellitus in Germany from 2010 until 2040. \u003cem\u003ePopul. health metrics\u003c/em\u003e. \u003cb\u003e22\u003c/b\u003e \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1186/s12963-024-00337-x\u003c/span\u003e\u003cspan address=\"10.1186/s12963-024-00337-x\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2024).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eED, P. et al. Economic Costs of Diabetes in the U.S. in 2022. \u003cem\u003eDiabetes care\u003c/em\u003e. \u003cb\u003e47\u003c/b\u003e \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.2337/dci23-0085\u003c/span\u003e\u003cspan address=\"10.2337/dci23-0085\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2024).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003e2. Diagnosis and Classification of Diabetes: Standards of Care in Diabetes-2024. \u003cem\u003eDiabetes care\u003c/em\u003e. \u003cb\u003e47\u003c/b\u003e \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.2337/dc24-S002\u003c/span\u003e\u003cspan address=\"10.2337/dc24-S002\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2024).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eDludla, P. V. et al. Pancreatic β-cell dysfunction in type 2 diabetes: Implications of inflammation and oxidative stress. \u003cem\u003eWorld J. Diabetes\u003c/em\u003e. \u003cb\u003e14\u003c/b\u003e \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.4239/wjd.v14.i3.130\u003c/span\u003e\u003cspan address=\"10.4239/wjd.v14.i3.130\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2025).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMezza, T. et al. β-Cell Fate in Human Insulin Resistance and Type 2 Diabetes: A Perspective on Islet Plasticity. \u003cem\u003eDiabetes\u003c/em\u003e \u003cb\u003e68\u003c/b\u003e, 1121. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.2337/db18-0856\u003c/span\u003e\u003cspan address=\"10.2337/db18-0856\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2025).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eA, D. et al. Pharmacotherapy of type 2 diabetes: An update and future directions. \u003cem\u003eMetab. Clin. Exp.\u003c/em\u003e \u003cb\u003e137\u003c/b\u003e \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1016/j.metabol.2022.155332\u003c/span\u003e\u003cspan address=\"10.1016/j.metabol.2022.155332\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2022).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKM, P. et al. Clinical characteristics, complications, comorbidities and treatment patterns among patients with type 2 diabetes mellitus in a large integrated health system. \u003cem\u003eBMJ open. diabetes Res. care\u003c/em\u003e. \u003cb\u003e3\u003c/b\u003e \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1136/bmjdrc-2015-000093\u003c/span\u003e\u003cspan address=\"10.1136/bmjdrc-2015-000093\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2015).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKI, G. et al. Interventions for Reversing Prediabetes: A Systematic Review and Meta-Analysis. \u003cem\u003eAm. J. Prev. Med.\u003c/em\u003e \u003cb\u003e62\u003c/b\u003e \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1016/j.amepre.2021.10.020\u003c/span\u003e\u003cspan address=\"10.1016/j.amepre.2021.10.020\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2022).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBorges-Canha, M. et al. Prediabetes remission after bariatric surgery: a 4-years follow-up study. \u003cem\u003eBMC Endocr. Disorders\u003c/em\u003e. \u003cb\u003e24\u003c/b\u003e \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1186/s12902-024-01537-0\u003c/span\u003e\u003cspan address=\"10.1186/s12902-024-01537-0\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2025).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eH, Z. et al. Exercise training modalities in prediabetes: a systematic review and network meta-analysis. \u003cem\u003eFront. Endocrinol.\u003c/em\u003e \u003cb\u003e15\u003c/b\u003e \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.3389/fendo.2024.1308959\u003c/span\u003e\u003cspan address=\"10.3389/fendo.2024.1308959\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2024).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eP, M., Z, S. H., F, A. \u0026amp; B. \u0026amp; Dietary pattern scores in relation to pre-diabetes regression to normal glycemia or progression to type 2 diabetes: a 9-year follow-up. \u003cem\u003eBMC Endocr. disorders\u003c/em\u003e. \u003cb\u003e23\u003c/b\u003e \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1186/s12902-023-01275-9\u003c/span\u003e\u003cspan address=\"10.1186/s12902-023-01275-9\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2023).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eT, S. et al. Quantifying the Relationship Between Physical Activity Energy Expenditure and Incident Type 2 Diabetes: A Prospective Cohort Study of Device-Measured Activity in 90,096 Adults. \u003cem\u003eDiabetes care\u003c/em\u003e. \u003cb\u003e46\u003c/b\u003e \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.2337/dc22-1467\u003c/span\u003e\u003cspan address=\"10.2337/dc22-1467\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2023).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eW, Y. et al. Different levels of physical activity and risk of developing type 2 diabetes among adults with prediabetes: a population-based cohort study. \u003cem\u003eNutr. J.\u003c/em\u003e \u003cb\u003e23\u003c/b\u003e \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1186/s12937-024-01013-4\u003c/span\u003e\u003cspan address=\"10.1186/s12937-024-01013-4\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2024).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eJA, F., BR, H., JK, R. J. P., LS, N. \u0026amp; S. \u0026amp; Changes in Cognition and Brain Function After 26 Weeks of Progressive Resistance Training in Older Adults at Risk for Diabetes: A Pilot Randomized Controlled Trial. \u003cem\u003eCan. J. diabetes\u003c/em\u003e. \u003cb\u003e47\u003c/b\u003e \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1016/j.jcjd.2023.01.004\u003c/span\u003e\u003cspan address=\"10.1016/j.jcjd.2023.01.004\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2023).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eM, L., MV, M. \u0026amp; A, K. \u0026amp; The effects of exercise treatment on learning and memory ability, and cognitive performance in diet-induced prediabetes animals. \u003cem\u003eSci. Rep.\u003c/em\u003e \u003cb\u003e10\u003c/b\u003e \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1038/s41598-020-72098-0\u003c/span\u003e\u003cspan address=\"10.1038/s41598-020-72098-0\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2020).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHH, L., C, Y. Z., ZJ, G. \u0026amp; C. \u0026amp; Meta-analysis of the effect of exercise intervention on cognitive function in elderly patients with type 2 diabetes mellitus. \u003cem\u003eBMC Geriatr.\u003c/em\u003e \u003cb\u003e24\u003c/b\u003e \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1186/s12877-024-05352-z\u003c/span\u003e\u003cspan address=\"10.1186/s12877-024-05352-z\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2024).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eJW, S., AR, H. \u0026amp; SK, H. Greater Lower Limb Fatigability in People with Prediabetes than Controls. \u003cem\u003eMed. Sci. Sports. Exerc.\u003c/em\u003e \u003cb\u003e52\u003c/b\u003e \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1249/MSS.0000000000002238\u003c/span\u003e\u003cspan address=\"10.1249/MSS.0000000000002238\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2020).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eM, D. \u0026amp; A, K. Investigating the Effects of Diet-Induced Prediabetes on Skeletal Muscle Strength in Male Sprague Dawley Rats. \u003cem\u003eInt. J. Mol. Sci.\u003c/em\u003e \u003cb\u003e25\u003c/b\u003e \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.3390/ijms25074076\u003c/span\u003e\u003cspan address=\"10.3390/ijms25074076\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2024).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eEM, R. et al. Reduced skeletal muscle phosphocreatine concentration in type 2 diabetic patients: a quantitative image-based phosphorus-31 MR spectroscopy study. \u003cem\u003eAm. J. Physiol. Endocrinol. Metab.\u003c/em\u003e \u003cb\u003e315\u003c/b\u003e \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1152/ajpendo.00426.2017\u003c/span\u003e\u003cspan address=\"10.1152/ajpendo.00426.2017\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2018).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eE, F. et al. Insulin Resistance Is Associated With Reduced Mitochondrial Oxidative Capacity Measured by 31P-Magnetic Resonance Spectroscopy in Participants Without Diabetes From the Baltimore Longitudinal Study of Aging. \u003cem\u003eDiabetes\u003c/em\u003e \u003cb\u003e66\u003c/b\u003e \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.2337/db16-0754\u003c/span\u003e\u003cspan address=\"10.2337/db16-0754\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2017).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eA, H. et al. Human skeletal muscle mitochondrial dynamics in relation to oxidative capacity and insulin sensitivity. \u003cem\u003eDiabetologia\u003c/em\u003e \u003cb\u003e64\u003c/b\u003e \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1007/s00125-020-05335-w\u003c/span\u003e\u003cspan address=\"10.1007/s00125-020-05335-w\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2021).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eL, S. et al. The Response of Mitochondrial Respiration and Quantity in Skeletal Muscle and Adipose Tissue to Exercise in Humans with Prediabetes. \u003cem\u003eCells\u003c/em\u003e 10, (2021). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.3390/cells10113013\u003c/span\u003e\u003cspan address=\"10.3390/cells10113013\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eK, O. Skeletal muscle proteomics: current approaches, technical challenges and emerging techniques. \u003cem\u003eSkelet. Muscle\u003c/em\u003e. \u003cb\u003e1\u003c/b\u003e \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1186/2044-5040-1-6\u003c/span\u003e\u003cspan address=\"10.1186/2044-5040-1-6\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2011).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eDT, C., AS, D. \u0026amp; R, M.-J., JP, Q. \u0026amp; Mass spectrometry-based proteomics approaches to interrogate skeletal muscle adaptations to exercise. \u003cem\u003eScand. J. Med. Sci. Sports\u003c/em\u003e. \u003cb\u003e34\u003c/b\u003e \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1111/sms.14334\u003c/span\u003e\u003cspan address=\"10.1111/sms.14334\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2024).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eGonzalez-Freire, M. et al. The Human Skeletal Muscle Proteome Project: a reappraisal of the current literature. \u003cem\u003eJ. Cachexia Sarcopenia Muscle\u003c/em\u003e. \u003cb\u003e8\u003c/b\u003e (5). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1002/jcsm.12121\u003c/span\u003e\u003cspan address=\"10.1002/jcsm.12121\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2025).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSJ, H., BN, S., CA, S. \u0026amp; JG, B. The application of proteomics in muscle exercise physiology. \u003cem\u003eExpert Rev. Proteomics\u003c/em\u003e. \u003cb\u003e17\u003c/b\u003e \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1080/14789450.2020.1879647\u003c/span\u003e\u003cspan address=\"10.1080/14789450.2020.1879647\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2020).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eB, V. \u0026amp; G, D. J, O. d. B. On-Chip Comparison of the Performance of First- and Second-Generation Micropillar Array Columns. \u003cem\u003eAnalytical chemistry\u003c/em\u003e 95, (2023). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1021/acs.analchem.3c01829\u003c/span\u003e\u003cspan address=\"10.1021/acs.analchem.3c01829\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eStejskal, K. et al. Deep Proteome Profiling with Reduced Carryover Using Superficially Porous Microfabricated nanoLC Columns. \u003cem\u003eAnal. Chem.\u003c/em\u003e \u003cb\u003e94\u003c/b\u003e, 15930. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1021/acs.analchem.2c01196\u003c/span\u003e\u003cspan address=\"10.1021/acs.analchem.2c01196\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2025).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBarkovits, K. et al. Reproducibility, Specificity and Accuracy of Relative Quantification Using Spectral Library-based Data-independent Acquisition. \u003cem\u003eMol. Cell. Proteomics: MCP\u003c/em\u003e. \u003cb\u003e19\u003c/b\u003e, 181. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1074/mcp.RA119.001714\u003c/span\u003e\u003cspan address=\"10.1074/mcp.RA119.001714\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2025).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eK, F. et al. Data-Independent Acquisition: A Milestone and Prospect in Clinical Mass Spectrometry-Based Proteomics. \u003cem\u003eMol. Cell. proteomics: MCP\u003c/em\u003e. \u003cb\u003e23\u003c/b\u003e \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1016/j.mcpro.2024.100800\u003c/span\u003e\u003cspan address=\"10.1016/j.mcpro.2024.100800\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2024).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBC, S. et al. Generating high quality libraries for DIA MS with empirically corrected peptide predictions. \u003cem\u003eNat. Commun.\u003c/em\u003e \u003cb\u003e11\u003c/b\u003e \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1038/s41467-020-15346-1\u003c/span\u003e\u003cspan address=\"10.1038/s41467-020-15346-1\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2020).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eGovaert, E. et al. Comparison of fractionation proteomics for local SWATH library building. \u003cem\u003eProteomics\u003c/em\u003e \u003cb\u003e17\u003c/b\u003e, 1700052. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1002/pmic.201700052\u003c/span\u003e\u003cspan address=\"10.1002/pmic.201700052\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2025).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eR, L. et al. Benchmarking commonly used software suites and analysis workflows for DIA proteomics and phosphoproteomics. \u003cem\u003eNat. Commun.\u003c/em\u003e \u003cb\u003e14\u003c/b\u003e \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1038/s41467-022-35740-1\u003c/span\u003e\u003cspan address=\"10.1038/s41467-022-35740-1\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2023).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eL, S. et al. Metabolomic Profile of Skeletal Muscle and Its Change Under a Mixed-Mode Exercise Intervention in Progressively Dysglycemic Subjects. \u003cem\u003eFront. Endocrinol.\u003c/em\u003e \u003cb\u003e12\u003c/b\u003e \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.3389/fendo.2021.778442\u003c/span\u003e\u003cspan address=\"10.3389/fendo.2021.778442\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2021).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eRA, S. et al. Human skeletal muscle biopsy procedures using the modified Bergstr\u0026ouml;m technique. \u003cem\u003eJ. visualized experiments: JoVE\u003c/em\u003e. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.3791/51812\u003c/span\u003e\u003cspan address=\"10.3791/51812\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2014).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eIR, L., ON, V. S., RR, S. \u0026amp; J. \u0026amp; Quantitative assessment of in-solution digestion efficiency identifies optimal protocols for unbiased protein analysis. \u003cem\u003eMol. Cell. proteomics: MCP\u003c/em\u003e. \u003cb\u003e12\u003c/b\u003e \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1074/mcp.M112.025585\u003c/span\u003e\u003cspan address=\"10.1074/mcp.M112.025585\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2013).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBC, S. et al. Chromatogram libraries improve peptide detection and quantification by data independent acquisition mass spectrometry. \u003cem\u003eNat. Commun.\u003c/em\u003e \u003cb\u003e9\u003c/b\u003e \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1038/s41467-018-07454-w\u003c/span\u003e\u003cspan address=\"10.1038/s41467-018-07454-w\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2018).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLK, P., SC, J., MJ, M. \u0026amp; BC, S. Acquiring and Analyzing Data Independent Acquisition Proteomics Experiments without Spectrum Libraries. \u003cem\u003eMol. Cell. proteomics: MCP\u003c/em\u003e. \u003cb\u003e19\u003c/b\u003e \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1074/mcp.P119.001913\u003c/span\u003e\u003cspan address=\"10.1074/mcp.P119.001913\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2020).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eD, A. et al. Improving Precursor Selectivity in Data-Independent Acquisition Using Overlapping Windows. \u003cem\u003eJ. Am. Soc. Mass. Spectrom.\u003c/em\u003e \u003cb\u003e30\u003c/b\u003e \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1007/s13361-018-2122-8\u003c/span\u003e\u003cspan address=\"10.1007/s13361-018-2122-8\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2019).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eR, B. et al. Optimization of Experimental Parameters in Data-Independent Mass Spectrometry Significantly Increases Depth and Reproducibility of Results. \u003cem\u003eMol. Cell. proteomics: MCP\u003c/em\u003e. \u003cb\u003e16\u003c/b\u003e \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1074/mcp.RA117.000314\u003c/span\u003e\u003cspan address=\"10.1074/mcp.RA117.000314\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2017).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eA, K., S, T. \u0026amp; J, G., J, P. \u0026amp; Causal analysis approaches in Ingenuity Pathway Analysis. \u003cem\u003eBioinf. (Oxford England)\u003c/em\u003e. \u003cb\u003e30\u003c/b\u003e \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1093/bioinformatics/btt703\u003c/span\u003e\u003cspan address=\"10.1093/bioinformatics/btt703\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2014).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eD, S. et al. The STRING database in 2023: protein-protein association networks and functional enrichment analyses for any sequenced genome of interest. \u003cem\u003eNucleic Acids Res.\u003c/em\u003e \u003cb\u003e51\u003c/b\u003e \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1093/nar/gkac1000\u003c/span\u003e\u003cspan address=\"10.1093/nar/gkac1000\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2023).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eRozing, G. Micropillar array columns for advancing nanoflow HPLC. \u003cem\u003eMicrochem. J.\u003c/em\u003e 170. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttp://dx.doi.org/10.1016/j.microc.2021.106629\u003c/span\u003e\u003cspan address=\"10.1016/j.microc.2021.106629\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2021).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMatzinger, M. et al. Micropillar arrays, wide window acquisition and AI-based data analysis improve comprehensiveness in multiple proteomic applications. \u003cem\u003eNat. Commun.\u003c/em\u003e \u003cb\u003e15\u003c/b\u003e, 1019. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1038/s41467-024-45391-z\u003c/span\u003e\u003cspan address=\"10.1038/s41467-024-45391-z\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2025).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eK, S., K, M. \u0026amp; J, O. d. B., G, D., P, J. \u0026amp; Ultrasensitive NanoLC-MS of Subnanogram Protein Samples Using Second Generation Micropillar Array LC Technology with Orbitrap Exploris 480 and FAIMS PRO. \u003cem\u003eAnal. Chem.\u003c/em\u003e \u003cb\u003e93\u003c/b\u003e \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1021/acs.analchem.1c00990\u003c/span\u003e\u003cspan address=\"10.1021/acs.analchem.1c00990\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2021).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBG, C., Y, M. \u0026amp; P, J., M, G., S, G. \u0026amp; Using micro pillar array columns (\u0026micro;PAC) for the analysis of permethylated glycans. \u003cem\u003eAnalyst\u003c/em\u003e \u003cb\u003e146\u003c/b\u003e \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1039/d1an00643f\u003c/span\u003e\u003cspan address=\"10.1039/d1an00643f\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2021).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBerg, H. E. et al. Micro-pillar array columns (\u0026micro;PAC): An efficient tool for comparing tissue and cultured cells of glioblastoma. \u003cem\u003eJ. Chromatogr. Open.\u003c/em\u003e \u003cb\u003e2\u003c/b\u003e, 100047 (2025).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eB, V. \u0026amp; G, D. J, O. d. B. Column-Only Band Broadening in a Porous Shell Radially Elongated Pillar Array Column. \u003cem\u003eAnalytical chemistry\u003c/em\u003e 96, (2024). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1021/acs.analchem.3c05756\u003c/span\u003e\u003cspan address=\"10.1021/acs.analchem.3c05756\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eJ, D. \u0026amp; P, L. C, B., A, S. Isolation Window Optimization of Data-Independent Acquisition Using Predicted Libraries for Deep and Accurate Proteome Profiling. \u003cem\u003eAnalytical chemistry\u003c/em\u003e 92, (2020). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1021/acs.analchem.0c00994\u003c/span\u003e\u003cspan address=\"10.1021/acs.analchem.0c00994\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eTS, B. \u0026amp; JV, O. Offline High pH Reversed-Phase Peptide Fractionation for Deep Phosphoproteome Coverage. \u003cem\u003eMethods Mol. biology (Clifton N J)\u003c/em\u003e. \u003cb\u003e1355\u003c/b\u003e \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1007/978-1-4939-3049-4_12\u003c/span\u003e\u003cspan address=\"10.1007/978-1-4939-3049-4_12\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2016).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eF, Y., DG, Y. S., RD, S. \u0026amp; C. \u0026amp; High-pH reversed-phase chromatography with fraction concatenation for 2D proteomic analysis. \u003cem\u003eExpert Rev. Proteomics\u003c/em\u003e. \u003cb\u003e9\u003c/b\u003e \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1586/epr.12.15\u003c/span\u003e\u003cspan address=\"10.1586/epr.12.15\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2012).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eD, Y. et al. Separation Orthogonality in Liquid Chromatography-Mass Spectrometry for Proteomic Applications: Comparison of 16 Different Two-Dimensional Combinations. \u003cem\u003eAnal. Chem.\u003c/em\u003e \u003cb\u003e92\u003c/b\u003e \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1021/acs.analchem.9b05407\u003c/span\u003e\u003cspan address=\"10.1021/acs.analchem.9b05407\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2020).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eJ, P., GN, M. A. \u0026amp; JA, S. BC, C. A gas phase fractionation acquisition scheme integrating ion mobility for rapid diaPASEF library generation. \u003cem\u003eProteomics\u003c/em\u003e \u003cb\u003e23\u003c/b\u003e \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1002/pmic.202200038\u003c/span\u003e\u003cspan address=\"10.1002/pmic.202200038\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2023).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eRB, K., JC, Y. \u0026amp; YJ, C. Advances in data-independent acquisition mass spectrometry towards comprehensive digital proteome landscape. \u003cem\u003eMass Spectrom. Rev.\u003c/em\u003e \u003cb\u003e42\u003c/b\u003e \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1002/mas.21781\u003c/span\u003e\u003cspan address=\"10.1002/mas.21781\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2023).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eC, G. et al. Extensive and Accurate Benchmarking of DIA Acquisition Methods and Software Tools Using a Complex Proteomic Standard. \u003cem\u003eJ. Proteome Res.\u003c/em\u003e \u003cb\u003e20\u003c/b\u003e \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1021/acs.jproteome.1c00490\u003c/span\u003e\u003cspan address=\"10.1021/acs.jproteome.1c00490\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2021).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eCA, S. et al. Slow-twitch fiber proportion in skeletal muscle correlates with insulin responsiveness. \u003cem\u003eJ. Clin. Endocrinol. Metab.\u003c/em\u003e \u003cb\u003e98\u003c/b\u003e \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1210/jc.2012-3876\u003c/span\u003e\u003cspan address=\"10.1210/jc.2012-3876\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2013).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003ePH, A. et al. Human muscle fiber type-specific insulin signaling: impact of obesity and type 2 diabetes. \u003cem\u003eDiabetes\u003c/em\u003e \u003cb\u003e64\u003c/b\u003e \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.2337/db14-0590\u003c/span\u003e\u003cspan address=\"10.2337/db14-0590\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2015).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eM, M. et al. Protein profile of fiber types in human skeletal muscle: a single-fiber proteomics study. \u003cem\u003eSkelet. Muscle\u003c/em\u003e. \u003cb\u003e11\u003c/b\u003e \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1186/s13395-021-00279-0\u003c/span\u003e\u003cspan address=\"10.1186/s13395-021-00279-0\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2021).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eC, L. et al. Repeated sprint training in hypoxia induces specific skeletal muscle adaptations through S100A protein signaling. \u003cem\u003eFASEB journal: official publication Federation Am. Soc. Experimental Biology\u003c/em\u003e. \u003cb\u003e38\u003c/b\u003e \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1096/fj.202302084RR\u003c/span\u003e\u003cspan address=\"10.1096/fj.202302084RR\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2024).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eP, S., KF, P., GI, S. \u0026amp; JW, K. Mitochondrial Dysfunction, Insulin Resistance, and Potential Genetic Implications. \u003cem\u003eEndocrinology\u003c/em\u003e 161, (2020). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1210/endocr/bqaa017\u003c/span\u003e\u003cspan address=\"10.1210/endocr/bqaa017\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eGN, R., AL, C. \u0026amp; TM, C. KS, N. Altered mitochondrial function in insulin-deficient and insulin-resistant states. \u003cem\u003eJ. Clin. Investig.\u003c/em\u003e \u003cb\u003e128\u003c/b\u003e \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1172/JCI120843\u003c/span\u003e\u003cspan address=\"10.1172/JCI120843\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2018).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eCB, N. Interplay between lipids and branched-chain amino acids in development of insulin resistance. \u003cem\u003eCell Metabol.\u003c/em\u003e \u003cb\u003e15\u003c/b\u003e \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1016/j.cmet.2012.01.024\u003c/span\u003e\u003cspan address=\"10.1016/j.cmet.2012.01.024\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2012).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eH, Z. et al. Branched-chain amino acid supplementation impairs insulin sensitivity and promotes lipogenesis during exercise in diet-induced obese mice. \u003cem\u003eObes. (Silver Spring Md)\u003c/em\u003e. \u003cb\u003e30\u003c/b\u003e \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1002/oby.23394\u003c/span\u003e\u003cspan address=\"10.1002/oby.23394\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2022).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eC, W., WJ, G. X. S. A. T., CT, L. \u0026amp; F. \u0026amp; Transcriptional profiles of type 2 diabetes in human skeletal muscle reveal insulin resistance, metabolic defects, apoptosis, and molecular signatures of immune activation in response to infections. \u003cem\u003eBiochem. Biophys. Res. Commun.\u003c/em\u003e \u003cb\u003e482\u003c/b\u003e \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1016/j.bbrc.2016.11.055\u003c/span\u003e\u003cspan address=\"10.1016/j.bbrc.2016.11.055\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2017).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eRA, A. \u0026amp; V, S. Role of complement in diabetes. \u003cem\u003eMol. Immunol.\u003c/em\u003e \u003cb\u003e114\u003c/b\u003e \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1016/j.molimm.2019.07.031\u003c/span\u003e\u003cspan address=\"10.1016/j.molimm.2019.07.031\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2019).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eH, E. Influence of Obesity and Type 2 Diabetes on Calcium Handling by Skeletal Muscle: Spotlight on the Sarcoplasmic Reticulum and Mitochondria. \u003cem\u003eFront. Physiol.\u003c/em\u003e \u003cb\u003e12\u003c/b\u003e \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.3389/fphys.2021.758316\u003c/span\u003e\u003cspan address=\"10.3389/fphys.2021.758316\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2021).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eG, S., GR, S. V., H, Z. \u0026amp; M., V, G. \u0026amp; Altered calcium homeostasis is correlated with abnormalities of fasting serum glucose, insulin resistance, and beta-cell function in the Newfoundland population. \u003cem\u003eDiabetes\u003c/em\u003e \u003cb\u003e54\u003c/b\u003e \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.2337/diabetes.54.11.3336\u003c/span\u003e\u003cspan address=\"10.2337/diabetes.54.11.3336\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2005).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eH, D. et al. Advanced glycation end products induce skeletal muscle atrophy and insulin resistance via activating ROS-mediated ER stress PERK/FOXO1 signaling. \u003cem\u003eAm. J. Physiol. Endocrinol. Metab.\u003c/em\u003e \u003cb\u003e324\u003c/b\u003e \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1152/ajpendo.00218.2022\u003c/span\u003e\u003cspan address=\"10.1152/ajpendo.00218.2022\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2023).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eZ, L. et al. Vitamin B6 Prevents Endothelial Dysfunction, Insulin Resistance, and Hepatic Lipid Accumulation in Apoe (-/-) Mice Fed with High-Fat Diet. \u003cem\u003eJournal of diabetes research\u003c/em\u003e (2016). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1155/2016/1748065\u003c/span\u003e\u003cspan address=\"10.1155/2016/1748065\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2016).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eE, M. \u0026amp; F, V. Vitamin B6 and Diabetes: Relationship and Molecular Mechanisms. \u003cem\u003eInt. J. Mol. Sci.\u003c/em\u003e \u003cb\u003e21\u003c/b\u003e \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.3390/ijms21103669\u003c/span\u003e\u003cspan address=\"10.3390/ijms21103669\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2020).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eF, C. et al. Exercise activates autophagy and regulates endoplasmic reticulum stress in muscle of high-fat diet mice to alleviate insulin resistance. \u003cem\u003eBiochem. Biophys. Res. Commun.\u003c/em\u003e \u003cb\u003e601\u003c/b\u003e \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1016/j.bbrc.2022.02.058\u003c/span\u003e\u003cspan address=\"10.1016/j.bbrc.2022.02.058\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2022).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eM, F., F, F. \u0026amp; E, H., P, F. \u0026amp; New insights into ER stress-induced insulin resistance. \u003cem\u003eTrends Endocrinol. Metab.\u003c/em\u003e \u003cb\u003e23\u003c/b\u003e \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1016/j.tem.2012.06.003\u003c/span\u003e\u003cspan address=\"10.1016/j.tem.2012.06.003\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2012).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eY. The PRIDE database resources in 2022: a hub for mass spectrometry-based proteomics evidences. \u003cem\u003eNucleic Acids Res.\u003c/em\u003e \u003cb\u003e50\u003c/b\u003e \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1093/nar/gkab1038\u003c/span\u003e\u003cspan address=\"10.1093/nar/gkab1038\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2022).\u003c/span\u003e\u003c/li\u003e\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":false,"highlight":"","institution":"","isAcceptedByJournal":true,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"scientific-reports","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"scirep","sideBox":"Learn more about [Scientific Reports](http://www.nature.com/srep/)","snPcode":"","submissionUrl":"","title":"Scientific Reports","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"stoa","reportingPortfolio":"Scientific Reports","inReviewEnabled":true,"inReviewRevisionsEnabled":true},"keywords":"","lastPublishedDoi":"10.21203/rs.3.rs-6331082/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-6331082/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003ePhysical exercise of even a moderate intensity is beneficial in both the prevention of prediabetes and management of Type 2 diabetes mellitus, as skeletal muscle is a primary tissue responsible for glucose uptake. Exercise-evoked proteomic alterations in muscle of subjects with prediabetes are of great importance for the study of relationships between insulin resistance and exercise. Due to its molecular composition proteomic analysis of skeletal muscle is challenging. To identify optimum approach, we compared various ion-chromatogram libraries assembled with the use of off-line high-pH fractionation (HpH), gas-phase fractionation (GPF) and libraryless DirectDIA\u0026trade; in LC/MS/HRMS DIA proteomic analysis of muscle from normoglycemic (NGT) and prediabetic (IGT) subjects after 3 months of supervised, mixed-mode exercise. GPF-fractionated, hybrid DDA/DIA libraries yielded the best overall balance between the speed of preparation, data collection and protein identification. Analysis revealed, that despite 3-month exercise intervention skeletal muscle from IGT subjects displayed significant alterations in pathways and molecules relevant to muscle contraction, extracellular matrix composition and protein synthesis as compared to NGT counterparts. In conclusion, our study underlines the importance of the selection of appropriate approach in the analysis of challenging clinical samples and reveals the potential explanation for deficiency of muscle function in the prediabetic state.\u003c/p\u003e","manuscriptTitle":"Ion-chromatogram libraries assembly in DIA proteomic analysis of post-exercise skeletal muscle in prediabetic subjects","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2025-04-23 06:36:08","doi":"10.21203/rs.3.rs-6331082/v1","editorialEvents":[{"type":"communityComments","content":0},{"type":"decision","content":"Revision requested","date":"2025-06-13T04:52:55+00:00","index":"","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2025-06-12T07:31:50+00:00","index":"hide","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2025-06-10T10:50:28+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"115749783825117529892993801070896048632","date":"2025-06-07T11:20:22+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"152028465440103851020627294363832762609","date":"2025-06-04T06:03:39+00:00","index":"hide","fulltext":""},{"type":"reviewersInvited","content":"","date":"2025-04-02T08:28:56+00:00","index":"","fulltext":""},{"type":"editorAssigned","content":"","date":"2025-04-02T08:22:07+00:00","index":"","fulltext":""},{"type":"editorInvited","content":"","date":"2025-04-02T04:40:09+00:00","index":"","fulltext":""},{"type":"checksComplete","content":"","date":"2025-03-31T12:16:07+00:00","index":"","fulltext":""},{"type":"submitted","content":"Scientific Reports","date":"2025-03-28T23:16:14+00:00","index":"","fulltext":""}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"scientific-reports","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"scirep","sideBox":"Learn more about [Scientific Reports](http://www.nature.com/srep/)","snPcode":"","submissionUrl":"","title":"Scientific Reports","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"stoa","reportingPortfolio":"Scientific Reports","inReviewEnabled":true,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"903aa75a-7819-41ab-8839-f99f1b7ada19","owner":[],"postedDate":"April 23rd, 2025","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"under-review","subjectAreas":[{"id":46997951,"name":"Biological sciences/Biochemistry/Proteomics"},{"id":46997952,"name":"Health sciences/Endocrinology/Endocrine system and metabolic diseases"},{"id":46997953,"name":"Biological sciences/Biochemistry/Hormones"},{"id":46997954,"name":"Health sciences/Diseases/Endocrine system and metabolic diseases/Diabetes/Type 2 diabetes mellitus"},{"id":46997955,"name":"Health sciences/Diseases/Endocrine system and metabolic diseases/Metabolic syndrome"},{"id":46997956,"name":"Health sciences/Diseases/Endocrine system and metabolic diseases/Obesity"},{"id":46997957,"name":"Health sciences/Diseases/Endocrine system and metabolic diseases/Pre diabetes"},{"id":46997958,"name":"Health sciences/Anatomy/Musculoskeletal system/Muscle/Skeletal muscle"},{"id":46997959,"name":"Health sciences/Biomarkers/Predictive markers"},{"id":46997960,"name":"Biological sciences/Physiology/Metabolism/Metabolic diseases/Diabetes/Type 2 diabetes mellitus"},{"id":46997961,"name":"Physical sciences/Chemistry/Analytical chemistry/Mass spectrometry"},{"id":46997962,"name":"Physical sciences/Chemistry/Analytical chemistry/Medical and clinical diagnostics"},{"id":46997963,"name":"Biological sciences/Systems biology/Biochemical networks"},{"id":46997964,"name":"Biological sciences/Molecular biology/Proteomics"}],"tags":[],"updatedAt":"2025-07-28T10:08:35+00:00","versionOfRecord":[],"versionCreatedAt":"2025-04-23 06:36:08","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-6331082","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-6331082","identity":"rs-6331082","version":["v1"]},"buildId":"8U1c8b4HqxoKbykW_rLl7","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}