diffONT: predicting methylation-specific PCR biomarkers based on nanopore sequencing data for clinical application

diffONT: predicting methylation-specific PCR biomarkers based on nanopore sequencing data for clinical application | bioRxiv /* */ /* */ <!-- <!-- /*! * yepnope1.5.4 * (c) WTFPL, GPLv2 */ (function(a,b,c){function d(a){return"[object Function]"==o.call(a)}function e(a){return"string"==typeof a}function f(){}function g(a){return!a||"loaded"==a||"complete"==a||"uninitialized"==a}function h(){var a=p.shift();q=1,a?a.t?m(function(){("c"==a.t?B.injectCss:B.injectJs)(a.s,0,a.a,a.x,a.e,1)},0):(a(),h()):q=0}function i(a,c,d,e,f,i,j){function k(b){if(!o&&g(l.readyState)&&(u.r=o=1,!q&&h(),l.onload=l.onreadystatechange=null,b)){"img"!=a&&m(function(){t.removeChild(l)},50);for(var d in y[c])y[c].hasOwnProperty(d)&&y[c][d].onload()}}var j=j||B.errorTimeout,l=b.createElement(a),o=0,r=0,u={t:d,s:c,e:f,a:i,x:j};1===y[c]&&(r=1,y[c]=[]),"object"==a?l.data=c:(l.src=c,l.type=a),l.width=l.height="0",l.onerror=l.onload=l.onreadystatechange=function(){k.call(this,r)},p.splice(e,0,u),"img"!=a&&(r||2===y[c]?(t.insertBefore(l,s?null:n),m(k,j)):y[c].push(l))}function j(a,b,c,d,f){return q=0,b=b||"j",e(a)?i("c"==b?v:u,a,b,this.i++,c,d,f):(p.splice(this.i++,0,a),1==p.length&&h()),this}function k(){var a=B;return a.loader={load:j,i:0},a}var l=b.documentElement,m=a.setTimeout,n=b.getElementsByTagName("script")[0],o={}.toString,p=[],q=0,r="MozAppearance"in l.style,s=r&&!!b.createRange().compareNode,t=s?l:n.parentNode,l=a.opera&&"[object Opera]"==o.call(a.opera),l=!!b.attachEvent&&!l,u=r?"object":l?"script":"img",v=l?"script":u,w=Array.isArray||function(a){return"[object Array]"==o.call(a)},x=[],y={},z={timeout:function(a,b){return b.length&&(a.timeout=b[0]),a}},A,B;B=function(a){function b(a){var a=a.split("!"),b=x.length,c=a.pop(),d=a.length,c={url:c,origUrl:c,prefixes:a},e,f,g;for(f=0;f<d;f++)g=a[f].split("="),(e=z[g.shift()])&&(c=e(c,g));for(f=0;f<b;f++)c=x[f](c);return c}function g(a,e,f,g,h){var i=b(a),j=i.autoCallback;i.url.split(".").pop().split("?").shift(),i.bypass||(e&&(e=d(e)?e:e[a]||e[g]||e[a.split("/").pop().split("?")[0]]),i.instead?i.instead(a,e,f,g,h):(y[i.url]?i.noexec=!0:y[i.url]=1,f.load(i.url,i.forceCSS||!i.forceJS&&"css"==i.url.split(".").pop().split("?").shift()?"c":c,i.noexec,i.attrs,i.timeout),(d(e)||d(j))&&f.load(function(){k(),e&&e(i.origUrl,h,g),j&&j(i.origUrl,h,g),y[i.url]=2})))}function h(a,b){function c(a,c){if(a){if(e(a))c||(j=function(){var a=[].slice.call(arguments);k.apply(this,a),l()}),g(a,j,b,0,h);else if(Object(a)===a)for(n in m=function(){var b=0,c;for(c in a)a.hasOwnProperty(c)&&b++;return b}(),a)a.hasOwnProperty(n)&&(!c&&!--m&&(d(j)?j=function(){var a=[].slice.call(arguments);k.apply(this,a),l()}:j[n]=function(a){return function(){var b=[].slice.call(arguments);a&&a.apply(this,b),l()}}(k[n])),g(a[n],j,b,n,h))}else!c&&l()}var h=!!a.test,i=a.load||a.both,j=a.callback||f,k=j,l=a.complete||f,m,n;c(h?a.yep:a.nope,!!i),i&&c(i)}var i,j,l=this.yepnope.loader;if(e(a))g(a,0,l,0);else if(w(a))for(i=0;i (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];var j=d.createElement(s);var dl=l!='dataLayer'?'&l='+l:'';j.src='//www.googletagmanager.com/gtm.js?id='+i+dl;j.type='text/javascript';j.async=true;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-M677548'); Skip to main content Home About Submit ALERTS / RSS Search for this keyword Advanced Search New Results diffONT: predicting methylation-specific PCR biomarkers based on nanopore sequencing data for clinical application View ORCID Profile Daria Meyer , View ORCID Profile Emanuel Barth , Laura Wiehle , Manja Marz doi: https://doi.org/10.1101/2025.02.17.638597 Daria Meyer 1 RNA Bioinformatics and High-Throughput Analysis, Friedrich Schiller University Jena , Leutragraben 1, 07743, Jena, Germany 2 Bioinformatics Core Facility, Friedrich Schiller University Jena , Leutragraben 1, 07743, Jena, Germany 3 oncgnostics GmbH , Löbstedter Str. 41, 07749, Jena, Germany Find this author on Google Scholar Find this author on PubMed Search for this author on this site ORCID record for Daria Meyer For correspondence: daria.meyer{at}uni-jena.de manja{at}uni-jena.de Emanuel Barth 2 Bioinformatics Core Facility, Friedrich Schiller University Jena , Leutragraben 1, 07743, Jena, Germany Find this author on Google Scholar Find this author on PubMed Search for this author on this site ORCID record for Emanuel Barth Laura Wiehle 3 oncgnostics GmbH , Löbstedter Str. 41, 07749, Jena, Germany Find this author on Google Scholar Find this author on PubMed Search for this author on this site Manja Marz 1 RNA Bioinformatics and High-Throughput Analysis, Friedrich Schiller University Jena , Leutragraben 1, 07743, Jena, Germany 2 Bioinformatics Core Facility, Friedrich Schiller University Jena , Leutragraben 1, 07743, Jena, Germany 4 Fritz Lipmann Institute, Leibniz Institute on Aging , , 07745, Jena, Germany 5 European Virus Bioinformatics Center , Leutragraben 1, 07743, Jena, Germany 6 Balance of the Microverse, cluster of Excellence , Grüne Aue, 07745, Jena, Germany Find this author on Google Scholar Find this author on PubMed Search for this author on this site For correspondence: daria.meyer{at}uni-jena.de manja{at}uni-jena.de Abstract Full Text Info/History Metrics Preview PDF Abstract DNA methylation is known to act as biomarker applicable for clinical diagnostics, especially in cancer detection. Methylation-specific PCR (MSP) is a widely used approach to screen patient samples fast and efficiently for differential methylation. During MSP, methylated regions are selectively amplified with specific primers. With nanopore sequencing, knowledge about DNA methylation is generated during direct DNA sequencing, without any need for pretreatment of the DNA. Multiple methods, mainly developed for whole-genome bisulfite sequencing (WGBS) data, exist to predict differentially methylated regions (DMRs) in the genome. However, the predicted DMRs are often very large, and not sufficiently discriminating to generate meaningful results in MSP creating a gap between theoretical cancer marker research and practical application, as no tool currently provides methylation difference predictions tailored for PCR-based diagnostics. Here we present diffONT , which predicts differentially methylated primer regions, directly suitable for MSP primer design and thus allowing a direct translation into practical approaches. diffONT takes into account (i) the specific length of primer and amplicon regions, (ii) the fact that one condition should be unmethylated, and (iii) a minimal required amount of differentially methylated cytosines within the primer regions. Based on two nanopore sequencing data sets we compared the results of diffONT to metilene , DSS and pycoMeth . We show that the regions predicted by diffONT are more specific towards hypermethylated regions and more usable for MSP. diffONT accelerates the design of methylation-specific diagnostic assays, bridging the gap between theoretical research and clinical application. Introduction For the detection of genome-wide DNA methylations different methods have been used over the last decades, including bisulfite sequencing, DNA digestion with methylation-sensitive restriction enzymes, and affinity purification of methylated DNA combined with array hybridization ( 1 ). Each method has its specific applications, advantages, and limitations ( 1 ). Bisulfite treatment ( 2 ), the gold standard for detecting methylated cytosines ( 3 ), offers high-resolution mapping of methylation sites but can be labor-intensive and costly. Furthermore, bisulfite treatment is harsh and requires a high amount of input DNA. Digestion with methylation-sensitive restriction enzymes provides a more straightforward approach but lacks the resolution and comprehensiveness of bisulfite sequencing. Affinity purification of methylated DNA, on the other hand, enables the enrichment of methylated regions but may introduce biases depending on the antibody or binding protein used ( 4 ). Besides, it only indirectly resolves methylation patterns. The choice of method often depends on the specific research question, sample availability, and required resolution. In cancer research, the accurate detection of DNA methylation is crucial for understanding tumorigenesis and for developing diagnostic and therapeutic strategies. One approach used in clinical diagnostic tests is methylation-specific PCR (MSP) ( 5 ), which uses bisulfite treatment, converting unmethylated cytosines to uracils while leaving methylated cytosines unchanged, allowing for the differentiation between methylated and unmethylated DNA by specific primers only binding to methylated sites. MSP is both simple and economical, making it suitable for population-wide screening ( 4 ). The design of MSP primers has to balance (a) specificity, to avoid creating unspecific PCR products or false positives in healthy individuals, and (b) efficiency, to achieve the theoretical maximum of duplication in each amplification step. Generally, a primer length of 18 to 24 nt is used, influencing specificity, with bisulfite-conversion-based PCR typically requiring longer primers (20 to 30 nt) due to lower sequence complexity( 6 ; 7 ). PCR products of 150 to 1000 bp are produced during amplification, with shorter products being more specific; clinical assays advise 120 to 300 bp, and PCR products based on bisulfite treated DNA are advised to be below 400 nt( 6 ; 7 ). The GC content of the PCR primers should be moderate, with 50-60% being ideal ( 8 ). For bisulfite-conversion-based PCR the primer sequence should also have enough non-CpG cytosines ( 5 ). Fig. 1 shows an example of an MSP primer pair design for a given DNA sequence. Download figure Open in new tab Fig. 1. Standard primer design workflow, assuming all CpG dinucleotides are completely methylated. By treating the DNA with sodium-bisulfite (bisulfite), all non-methylated cytosines are deaminated into uracils. In a following PCR, the methylation specific primers bind only to methylated DNA stretches. Thus, only methylated DNA stretches are amplified, unmethylated DNA is not amplified. This allows a methylation-specific amplification of DNA. Primer length: 25 nt, PCR product length: 120 nt. For the detection of marker regions for MSP, ONT sequencing is advantageous as it considers the entire genome, providing a comprehensive, unprejudiced view beyond just the regions described in literature or known disease-associated genes. This method surpasses the limitations of focusing solely on known regions, offering a broader and more inclusive analysis, while also directly measuring DNA methylation without the need of an amplification or conversion step on a single nucleotide resolution ( 9 ). While nanopore sequencing is not applicable as a diagnostics means in most applications due to its lack of economy and sensitivity, it is well-suited for detecting differentially methylated regions using tools developed for short or long-read sequencing ( 10 ; 11 ; 12 ; 13 ). These regions could potentially serve as targets for MSP. A variety of bioinformatic tools are available for measuring DNA methylation differences, primarily designed for short reads or bisulfite data analysis: DSS (Dispersion Shrinkage for Sequencing data) ( 14 ) is specifically designed to accommodate a limited number of biological replicates. It has an integrated beta-binomial regression model with an ’arcsine’ link function, utilizing a Bayesian hierarchical framework for differential methylation analysis. It assesses per-cytosine significance using a Wald test. However, DSS has been noted in a comparative study for its slower computational speed and potential memory issues, as well as a tendency to identify regions that may exceed the actual size of DMRs ( 15 ). It has also been observed to have lower precision when dealing with small number of replicates, with a notable drop in precision observed for n=3 compared to n=10 ( 10 ; 16 ). metilene ( 11 ) is designed for high-speed performance and efficient memory usage ( 17 ; 11 ) to efficiently process large datasets. Additionally, it has been reported to work well with low coverage data. It has a binary segmentation algorithm implemented, coupled with a two-dimensional statistical test for detecting DMRs. While metilene is not effective for analyzing methylation differences in single-CpG analysis ( 16 ), it is known to identify DMR boundaries more accurately than DSS ( 16 ). However, it has been noted to identify numerous DMRs with low mean methylation differences between groups, although this parameter can be adjusted ( 15 ). Additionally, when applied to nanopore sequencing data, the data needs to be formatted to be in an adequate input format for metilene . pycoMeth ( 12 ) is to our knowledge the only tool designed for differential methylation analysis using nanopore sequencing data processed through Nanopolish . pycoMeth shows increased recall and precision in DMR testing, especially for detecting low effect-size methylation changes in low-coverage settings ( 12 ). Additionally, by utilizing Bayesian changepoint detection, pycoMeth accommodates methylation call uncertainties instead of binarization of methylation probabilities as the other tools described above ( 12 ). pycoMeth is suitable for comparing two or more samples. pycoMeth focuses on identifying methylated regions based on nanopore sequencing data. However, it fails to support newer methylation callers such as implemented in the current state-of-the-art basecaller Dorado 1 . The various analysis tools for identifying differential methylation patterns differ significantly in performance, including speed, memory usage, number of sample groups to compare, coverage, and the ability to detect small effect sizes. However, most of these tools are not suitable for the analysis of nanopore data. To date, no tool has been specifically designed for the adequate detection of primer binding sites or the prediction of potential biomarkers that can be easily and affordably detected using methylation-specific PCR (MSP). Especially, identifying DNA methylation marker regions that can be effectively translated into clinical practice remains a significant challenge ( 4 ). Biomarkers not only need to be clearly differentially methylated in disease but also must be suitable for clinical diagnostic tests, ensuring that healthy individuals are minimally misclassified as having cancer, by minimizing the number of false positive test results and at the same time maximizing detection rates of diseased individuals. Here, we present the workflow diffONT , written in Python, which addresses critical gaps in DNA methylation marker discovery and validation, particularly for clinical diagnostic via MSP. diffONT predicts potential diagnostic biomarkers based on low-coverage nanopore sequenced whole genome DNA methylation data independent of the used methylation caller, considering clinical constraints by minimizing false positives in the prediction of candidate regions and applying it to clinical datasets. diffONT allows for customization of parameters such as methylation thresholds, region length, minimum differential CpG sites, and PCR product size. Using an Oxford Nanopore Technology (ONT) dataset for validation, our tool has been benchmarked against established methods like metilene , DSS , and pycoMeth , demonstrating fewer regions detected, however, with more restrict methylation differences and direct applicability for MSP. This approach not only explores upstream methylation patterns but also considers methylation dynamics across diverse genomic contexts, thereby enhancing the quality and relevance of identified biomarkers for clinical applications. Results In this study, we demonstrate the functionality of diffONT , a tool designed to predict methylation-specific PCR biomarkers from Nanopore-sequenced DNA data, using a publicly available ONT dataset. We describe the identification of CpGs of interest, followed by the definition of potential primers and the resulting MSP regions. To validate the effectiveness of our approach, we benchmark diffONT against other tools with similar functionality, namely metilene , DSS , and pycoMeth . Non-uniform distribution of CpGs of interest In order to identify biomarker candidates for clinical diagnostics, we first determined the CpG sites of interest according to Fig. 9A-G . These have to be nearly unmethylated in control samples (less than 10 % methylation) and show significant hypermethylation in disease samples. For the ONT dataset analyzed, about 3 874 242 CpGs (13,4 % of CpGs in the reference genome) of the complete human genome are predicted to be differentially methylated, see Fig. 2A . The distribution of CpGs of interest among the chromosomes generally correlates with chromosome length ( SFig. S1B ), with more CpGs of interest on the longer chromosomes 1 to 5 and fewer on the shorter ones, 19 to 22. Surprisingly, chromosome X forms an exception: despite being similar in size to chromosome 7, it contains 31.7 % more CpGs of interest than expected. To understand the unexpected behavior of chromosome X, we calculated for each chromosome the absolute number of (i) genes, (ii) CpGs, and (iii) CpG islands ( SFig. S2A-C ), as well as these attributes normalized by chromosome length ( SFig. S2D-F ). While these factors do not explain the large number of CpGs of interest, it was notable that chromosome X had comparatively low coverage, see STab. S2, which might suggest an artifact. A possible explanation might be abnormalities in the chromosomal copy numbers in the male cell line COLO829 ( 21 ), including, a duplication of the X chromosome. To determine whether the CpGs of interest are clustered on the chromosome or evenly distributed, we applied a sliding window of 10 000 nt across the entire genome (for comparison, the average CpG island is 700 nt long). Our analysis revealed that most windows have a ratio of CpGs of interest to total CpGs below 25 %, with only a few windows showing a ratio above 70 % CpGs of interest, see Fig. 2B . A chromosomal distribution can be found in SFig. S3A and choosing different window sizes does not change the observed effect, see SFig. S3B . A scatter plot reveals several interesting windows with a high percentage of CpGs of interest and a substantial total number of CpGs within the windows, see Fig. 2C . Notably, these windows overlap with the genes HIC1, GATA2, and TSPOAP1. Among these, HIC1 stands out as a tumor suppressor gene and features a CpG island spanning its promoter region ( 18 ). Also, GATA2 and TSPOAP1 show notable associations with cancer: GATA2 is implicated in hematological malignancies and potentially influences tumor microenvironment interactions ( 19 ), while TSPOAP1, involved in mitochondrial function and stress responses may affect tumor cell metabolism and behavior, highlighting their relevance in the dataset comprising normal B lymphoblasts (COLO829BL) and melanoma fibroblasts (COLO829) ( 22 ; 20 ). The fact that a known tumor suppressor gene stands out among the analyzed windows, in relation to a high percentage of CpGs of interest highlights the potential of our approach to identify relevant biomarkers in cancer. Download figure Open in new tab Fig. 2. (A) Distribution of CpGs of interest across human chromosomes. As expected, the distribution largely correlates with chromosome length; however, chromosome X contains more CpGs of interest than anticipated. The relative amount of CpGs of interest compared to the total number of CpGs per chromosome vary less, with low fractions on chromosomes 17 and 19 and a high fraction of CpGs of interest on chromosome 4. (B) Histogram showing the distribution of the ratio of CpGs of interest to total CpGs across the human genome. Ratios were calculated using 10 000 nt windows, shifted by half the window length, and binned into 0.05 increments. For ratios above 0.7, data was rebinned into 0.01 increments. The highest ratio (1.0) is on chromosome 9, with only one CpG of interest. The second highest (0.88) is on chromosome X, with seven CpGs of interest out of eight. (C) Scatter plot depicting the ratio of CpGs of interest to total CpGs for various windows, with the total number of CpGs indicated on the vertical axis. Outlier points with a high percentage of CpGs of interest and a high amount of CpGs for the window are marked and overlap the genes HIC1, GATA2, and TSPOAP1, which are known to be associated with cancer ( 18 ; 19 ; 20 ). CpG islands are regions of the genome rich in CpG dinucleotides and are often associated with gene promoters and regulatory regions ( 23 ). About 50 % of CpG islands contain known transcription start sites ( 24 ), particularly most housekeeping genes have CpG island covering their transcription start ( 25 ; 26 ). In our analysis, we found that, on average, only 5 % of the CpGs of interest overlap with these CpG islands, see Tab. 1 . Independent analysis of each chromosome reveals considerable variability, with overlaps ranging from 2.04 % on chromosome 4 to 16.79 % on chromosome 19. This variability does not align with the overall distribution of CpG islands across chromosomes, see SFig. S2C . Notably, some chromosomes exhibit a higher overlap than others, suggesting that specific regions might influence the presence of CpGs of interest. View this table: View inline View popup Download powerpoint Table 1. Overview of the percentage overlap between predicted CpGs of interest, primers, and MSP regions with CpG islands. The overlap increases with each pipeline step, from an average 5 % for CpGs to an average of 72 % for MSP regions. Methylation-specific PCR primer designed by diffONT For methylation-specific PCR, primers need to be 18-24 nt long, contain at least three CpGs of interest each, and receive a primer score reflecting the methylation difference between patient and control samples, see Fig. 9H/I . We provide a comprehensive list of potential primers, detailing their genomic position, length, number of CpGs of interest, score, and GC content. Applying diffONT to the ONT dataset resulted in 16 837 primer regions distributed across all human chromosomes. The majority of primers are predicted on chromosomes 1, X, 2, 17, and 6, with 1 425, 1 396, 1 261, 1 097, and 1 039 primers identified, respectively, see Fig. 3A . Most potential primers of diffONT (13 329) contain the required minimum of three CpGs of interest and only one primer contains eight CpGs of interest, see Fig. 3B . Tab. 2 provides a summary of the CpGs of interest compared to the total number of CpGs per primer. Overall, the number of CpGs of interest tends to increase as the total CpG count per primer rises. However, the majority of primers still contain only three CpGs of interest, regardless of the maximum number of CpGs. Over 96 % of 24 nt-wide windows (250 561 293 out of 258 326 583) do not contain any CpG of interest. While only about 5 % of CpGs of interest overlap with CpG islands, approximately 50 % of primers overlap with CpG islands, see Tab. 1 . It varies between 38.86 % on chromosome 10 and 61.38 % on chromosome 22. Download figure Open in new tab Fig. 3. (A) Distribution of primers predicted by diffONT across chromosomes. The distribution generally follows chromosome length ( SFig. S1 B ), with exceptions on chromosomes 17 and X, which have more CpGs of interest than expected. (B) In total, 16,837 primers contain at least three CpGs of interest, with the majority of primers predicted by diffONT containing exactly three CpGs of interest. Note: the y-axis is visualized on a logarithmic scale (log10). (C) Most individual primers predicted by diffONT show overlap. (D) The various primer combinations generate multiple MSP regions with slightly different scores and characteristics, such as length and number of CpGs, as shown in Fig. 9K . In CpG-rich regions, such as CpG islands, multiple overlapping primers are predicted ( Fig. 3C ). As multiple combinations of predicted primers are possible to yield a PCR product (within the range of 400 nt), multiple MSP regions containing the same primers are possible ( Fig. 3D ). The best MSP region is the combination of primers with the highest score (i.e. those primers with the highest coverage and non-control methylation). View this table: View inline View popup Download powerpoint Table 2. Distribution of primers (column 2) by the total number of CpGs (column 1) and the number of CpGs of interest (column 3). For example, among 1 342 primers containing 6 CpGs, 614 of those primers have 3 CpGs of interest, 434 have 4, 233 have 5, and 61 have all 6 CpGs of interest. Note that primers with a high number of total CpGs may be unsuitable due to elevated annealing temperatures. MSP regions by diffONT The best combination of primers is selected based on the highest summed score of two primers with a user-defined distance (60-400 nt), see Fig. 9J/K . diffONT presents a list of MSP regions containing score, coverage, and length. In total, the diffONT algorithm reports 6 745 potential MSP regions for the ONT dataset across the human genome. The highest number of MSP regions were predicted on chromosome X with 1 146 predicted MSP regions (see Fig. 4A ). The fewest hits are located on chromosomes 21, 18, and 13 with 72, 78, and 100 MSP regions, respectively. No suitable MSP region is predicted on chromosome Y, as the Y chromosome is lost in COLO829. The ONT dataset shows differences in the coverage between the chromosomes, see STab. S2, with e.g., very low coverage (0.27 – 0.4 X) on chromosome Y for the melanoma fibroblast samples. Download figure Open in new tab Fig. 4. (A) Distribution of predicted MSP regions by diffONT on the ONT dataset across chromosomes. The regions are unevenly distributed, with the highest concentrations on chromosomes X, 1, 2, and 17. (B) MSP regions overlap with (annotated) genomic features: 5 683 genes, 296 promoter regions (-500 nt), and 766 intergeneic regions. (C) Histogram of diffONT score values for all MSP regions predicted on the ONT dataset, with scores binned in intervals of 50. Two distinct distributions are observed: one peaking around 200 and another around 2 000, with scores above 5 000 appearing as outliers. (D) Visualization of MSP region scores versus mean methylation differences between melanoma fibroblast and normal B lymphoblast samples for all MSP regions predicted by diffONT on the ONT dataset. Color represents average coverage in melanoma fibroblast samples. Regions with low scores despite high methylation differences (bottom right) exhibit low coverage in melanoma fibroblast samples. However, the mean coverage of the predicted regions follows a similar bimodal distribution for both fibroblast and normal lymphoblast samples, see SFig. S1A . Most of the 6 745 MSP regions overlap with annotated genomic features, as shown in Fig. 4B . Specifically, 5 683 MSP regions (84 %) overlap with an annotated gene, another 296 regions (4.39 %) overlap with the 500 nt upstream of a gene and about 10 % (766 MSPs) are intergenic. The score for the 6 745 predicted MSP regions ranges from 16.44 to 8 978.00, see Fig. 4C . The score values follow a bimodal distribution, with a peak below 500, created by 1 504 MSP regions, and a second peak at around 2 000 created by 5 138 MSP regions with a score between 500 and 5 000. Only 101 MSP regions have a score above 5 000. The mean methylation difference values (also called effect size), generated by subtracting the average melanoma fibroblast samples’ methylation from the average normal B lymphoblast samples’ methylation, are shown in SFig. S4A reveals a high methylation difference (above 75 %) for about 50 % of the regions (3 306 out of 6 745). This is an interesting result, as the diffONT algorithm does not require high methylation differences. In contrast, only 317 of the predicted MSP regions have a mean methylation difference below 25 %. Comparing the MSP region score with the mean methylation difference for all MSP regions predicted by diffONT reveals a positive correlation, see Fig. 4D . Regions with a higher mean methylation difference tend to have higher scores, likely because the score is influenced by the average disease methylation, and regions with significant methylation differences typically have higher disease methylation, see SFig. S4B . However, some regions exhibit a high mean methylation difference but a low score, which can be attributed to low coverage in the melanoma fibroblast samples. Since the score incorporates both disease sample methylation and coverage, low coverage results in a lower score, explaining this discrepancy. A histogram of the average disease and control samples coverage shows a bimodal distribution of coverage values, with one peak at around 2 X coverage, and another one at around 15 X coverage, see SFig. S1A . We analyzed the overlap of the MSP regions predicted by diffONT with CpG islands, see Tab. 1 . On average, 72 % of MSP regions overlap at least one CpG island by at least one nucleotide. However, the overlap varies significantly across chromosomes: only 39 % and 42 % of MSP regions on chromosomes 14 and 15, respectively, overlap with CpG islands, while over 90 % of MSP regions overlap CpG islands on chromosomes 18, 21, 22, and X. Interestingly, while the average overlap with CpG islands across all chromosomes is about 5 % for individual CpGs of interest, it increases to 46 % for primer regions and to 72 % on MSP regions. The predicted region with the highest score overlaps an annotated CpG island as well as the 3’ UTR of the gene TSPAN33 (Tetraspanin 33), see Fig. 5 , a gene for which differential methylation is linked to several lymphomas ( 27 ). The methylation for the predicted region in the TSPAN33 3’ UTR strongly differs between the melanoma fibroblast and normal B lymphoblast replicates. Download figure Open in new tab Fig. 5. Visualization of the highest-scoring MSP region predicted by diffONT on the ONT dataset using the Integrative Genomics Viewer (IGV). The predicted diffONT region is annotated in blue at the top, with the gene TSPAN33 and the associated CpG island annotated below (blue and green, respectively). The 5mC methylation status is color-coded: red for methylated and blue for unmethylated sites. The melanoma fibroblast sample (5x) is displayed above the normal B lymphoblast sample (5x), showing strong methylation differences overlapping the TSPAN33 3’ UTR. Another region, which is among the 50 highest-scored MSP regions overlaps the 5’ UTR of transcription factor TCF20, see Fig. S4C and a predicted CpG island. Variants of TCF20 are known to impair neurodevelopmental disability ( 28 ). Additionally, TCF20 is known to be abundantly expressed in small cell lung cancer ( 29 ) and to distinguish desmoid tumors from nodular fasciitis ( 30 ). The predicted MSP region shows a higher amount of CpG dinucleotides than the surrounding genomic region and a high methylation difference between the fibroblast and lymphoblast samples. The top 10 ranked MSP regions can be found in STab. S3. Performance benchmark of diffONT To the best of our knowledge, none of the existing tools predicts MSP regions for PCRs in clinical applications. Instead, all currently available tools screen for differentially methylated regions (DMRs) in general. That is why we compared the MSP regions of our tool diffONT with DMRs predicted by metilene , DSS , and pycoMeth . We expect our MSP regions to form a subset of DMR regions, due to our stringent selections based on (i) an applicability in a clinical context, i.e., a high methylation in disease samples and almost no methylation in control samples, instead of finding general differences in methylation; and (ii) a high methylation rate within only 24 nt allowing primer design for methylation-specific PCR. Clinical nanopore sequencing data often has limited coverage (∼7 X per replicate per flow cell for the dataset used). We focus our benchmarking on metilene , which performs well with such data ( 11 ), and DSS , which is currently recommended by ONT for the analysis of nanopore sequencing data. Additionally, we show benchmarking data for pycoMeth , as this tool has been developed especially for DMR detection in nanopore sequencing data. Comparison to metilene metilene was implemented to study DNA methylation differences, e.g., after environmental stress in plants ( 31 ), during development ( 32 ; 33 ), and for the prediction of cancerous lung tissue ( 34 ). We applied metilene to predict DMRs to the same ONT dataset analyzed with diffONT . metilene predicted 621 523 DMRs, of which only 2 342 overlap with MSP regions of diffONT , see Fig. 6A . Contrarily, 6 032 (98 %) of 6 745 MSP regions predicted by diffONT overlap with a DMR predicted by metilene . There are mainly two differences between the regions predicted by diffONT and metilene : (i) the length of metilene ’s DMRs ranges from 10 nt to 7 486 nt, whereas diffONT ’s MSP regions range from 60 nt to 400 nt, see Fig. 6B , resulting in generally longer regions predicted by metilene , additionally, (ii) diffONT only considers up-regulated methylation patterns compared to control samples, whereas metilene predicts hyper- and hypomethylated regions ( Fig. 6C , green). These two differences explain the low fraction of DMRs not intersecting at all with MSP regions. Download figure Open in new tab Fig. 6. Comparison between diffONT (red) and metilene (green) for the ONT dataset, comparing cancerous fibroblast cells with normal lymphoblast cells. (A) Venn diagram showing the overlap of predicted regions. (B) Frequency density of the region length. All regions returned by diffONT range from 60 to 400 nt (vertical lines) due to the given values for minimum and maximum amplicon length in the algorithm, see Fig. 9J . Please note: x-axis is shown in log( 10 ) scale. (C) Frequency density of the mean methylation differences between the melanoma fibroblast and normal lymphoblast replicates. (D) The distribution of the four fields of the Venn diagram. (E) Number and (F) distribution of DMR (from metilene ) and MSP (from diffONT ) regions per chromosomes. The density distribution of mean methylation differences of the regions predicted by metilene ranges from -100 to 100. Interestingly, a peak around -1 may highlight regions relevant for studying subtle methylation changes, as needed e.g. in early disease stages. This is, however, not applicable for methylation-specific PCRs in a clinical context. The regions predicted by diffONT all have a methylation difference peak at around 93, thus diffONT focuses on regions with high methylation differences. Notably, for all the comparisons made, the diffONT score filter has not been applied at this stage, thus, applying it will further increase the rate of methylation differences among the finally selected MSP regions, leading to even more accurate results of diffONT . In order to understand the distribution of methylation differences in more detail, we further distinguished the four fields of the Venn Diagram from Fig. 6A in Fig. 6D . Due to the specific design of diffONT for clinical applicability, which limits the analysis to regions where control samples exhibit little to no methylation, most diffONT results display a very high mean methylation difference. Accordingly, the diffONT results can be better applied in the clinical context. The distribution, number, and appearance of diffONT and metilene regions across the chromosomes are reported in Tab. 3 , Fig. 6E (note the log-scaled x-axis) and Fig. 6F . Both tools predict some chromosomal parts to have no methylation differences, such as the 5’ end of acrocentric chromosomes 13, 14, 15, 21, and 22. View this table: View inline View popup Download powerpoint Table 3. Number of regions called per chromosome by diffONT and metilene on the ONT dataset. For both methods the distribution of predicted regions among the chromosomes is quite similar. Both methods predict most regions on chromosomes 1 and 2 and the least regions on chromosome 21. Low counts also exist on alternative chromosomes (data not shown). In terms of run time diffONT , though with a shorter pre-processing time, performs much slower than metilene . However, diffONT can predict highly methylated regions more precisely, see Tab. 4 . diffONT is computationally even more intensive for whole-genome analysis and more efficient for smaller regions, such as individual chromosomes. The longer runtime is initially counterintuitive given the smaller number of regions diffONT predicts. It arises, however, due to single nucleotide analysis and calculation of single nucleotide statistics ( Fig. 9E ). View this table: View inline View popup Download powerpoint Table 4. Comparison of the main result statistics between diffONT , metilene , and DSS on the ONT benchmarking dataset. Time calculated for pre-processing starting with (unsorted) bedmethyl files. # regions – number of DMRs and MSP regions, respectively. However, we can assume diffONT to predict hypermethylated regions more precisely than metilene , with higher accuracy. Comparison to DSS DSS is widely used in various publications for detecting differentially methylated regions (DMRs) based on nanopore sequencing data ( 35 ; 13 ; 36 ) and is also recommended by ONT. However, for whole genome data analysis, DSS is notably slow, see Tab. 4 . Therefore, in this work, DSS was applied to chromosome 17 only, which contains many regions predicted by both diffONT and metilene , see Fig. 6 . The runtime of DSS for only one chromosome exceeds that of diffONT by factor > 40. For the whole genome, we allowed DSS to run for over three months before halting the process without any result. DSS results of chromosome 17 could be filtered by p-value or mean methylation difference. The complete (unfiltered) number of reported DMRs was used for comparison. In total, DSS predicted 10 022 DMRs for the ONT dataset on chromosome 17, which is 20 times the number of regions predicted by diffONT (455) and about a quarter of the DMRs predicted by metilene (41 144), see Tab. 3 . The majority of diffONT regions (345 of 455, 76 %, Fig. 7A ) were found by both tools, another 107 (24 %) by metilene but not DSS , and only 2 regions by DSS but not metilene . Similar to metilene , also DSS detects hyper- and hypomethylated regions, which results in a high number of regions (ca. 7 500) found by both tools but not diffONT . This is inherent to diffONT , being designed to detect hypermethylated regions only, which additionally requires almost no methylation in control samples for the following application of MSP. The coverage of the chromosome for DMR and MSP regions is for diffONT 0.04 %, for metilene 14.45 %, and for DSS 2.28 %, see Fig. 7B . The DMRs predicted by DSS have a required minimum length of 51 nt, and no maximum length, resulting in DMRs ranging from 51 nt to 3 331 nt with a median region length of 122 nt. Nevertheless, DSS yields a high number (2 150) of long DMRs exceeding 250 nt. The region length distribution of DSS is similar to the one for regions predicted by metilene , with DMRs in the range of 10 nt to 3 576 nt and a median region length of 209 nt, Fig. 7C . Regions predicted by diffONT on the other hand have a minimum and maximum length defined (60 nt and 400 nt, respectively) and have a larger median region length of 182 nt. Expecting hypermethylation in the disease samples, diffONT screens exclusively for MSP regions with a positive mean methylation difference, Fig. 7D . In contrast, both DSS and metilene screen for DMRs with both positive and negative mean methylation differences. Similar to the entire human genome, metilene predicts numerous DMRs with a mean methylation difference near 0, whereas DSS predicts very few DMRs with small methylation differences. It can be summarized, that both DSS and metilene predict much more and longer regions on chromosome 17, compared to diffONT , however, for the application of MSP in the clinical context, diffONT seems more suitable. Download figure Open in new tab Fig. 7. Comparison of diffONT , metilene , and DSS on chromosome 17 using the ONT dataset. (A) Venn diagram of predicted regions found by diffONT , metilene , and DSS . The intersections contain numbers in different colors which refer to the different tools. For example, 7 976 DSS regions overlap with metilene regions, whereas only 7 386 metilene regions overlap with DSS regions. (B) Distribution of regions predicted by diffONT , metilene , and DSS among chromosome 17 and coverage of chromosome by the DMRs and MSP regions predicted by diffONT , metilene , and DSS . (C) Density distribution of the length of predicted regions by diffONT , metilene , and DSS . All regions returned by diffONT range from 60 to 400 nt (vertical lines) due to the given values for minimum and maximum amplicon length in the algorithm (see Fig. 9J ). Data is shown only for region length below 1 000 nt. (D) Density distribution of the mean methylation differences between the melanoma fibroblast and normal lymphoblast replicates for diffONT , metilene , and DSS . Comparison with DMR calling tool pycoMeth pycoMeth was the first tool being developed to predict DMRs from nanopore data, which includes storage, management and analysis of ONT DNA methylation data ( 12 ). Additionally, to our knowledge, pycoMeth is currently the only tool using low-coverage (up to 15 X) long-read nanopore sequencing data for predicting DMRs. For the comparison of diffONT to pycoMeth , both tools were applied to a male and a female sample, resulting in more than 100 000 predicted regions for each of the tools, Tab. 5 . View this table: View inline View popup Download powerpoint Table 5. Distribution of regions predicted by diffONT and pycoMeth on the GIAB dataset. In total, 104 805 and 157 660 regions were identified for diffONT and pycoMeth , respectively. No regions are predicted on chromosome Y, since female cells were compared to male cells. Low counts also exist on alternative chromosomes (data not shown). pycoMeth segments the genome based on methylation information into intervals, resulting in 584 171 intervals for the GIAB dataset. Differential methylation analysis of these intervals identified 157,660 intervals as DMRs, characterized by significant p-values. While about half of the MSP regions (54 122 out of 104 805) predicted by diffONT overlap a region predicted by pycoMeth , only a small fraction of regions predicted by pycoMeth (422 out of 157 660) are overlapping an MSP region, Fig. 8A . At first glance, this may appear to be a disproportionate distribution; however, it can be explained by the following observation: While the MSP regions predicted by diffONT have a defined length of 60-400 nt, the intervals predicted by pycoMeth range from 0 to 2 348 466 nt in length, see Fig. 8B , with a peak at 4 596 nt. Note, that DSS and metilene have not identified any such long DMR, see Fig. 6 , Download figure Open in new tab Fig. 8. Comparison between diffONT (red) and pycoMeth (purple) for the GIAB dataset. (A) Venn diagram showing the overlap of predicted regions. (B) Density distribution of the region length of predicted regions by diffONT and pycoMeth . All regions returned by diffONT range from 60-400 nt (vertical lines) due to the given values for minimum and maximum amplicon length in the algorithm, see Fig. 9J . The X-axis is shown in log( 10 ). The regions predicted by pycoMeth far exceed the length of those predicted by diffONT . (C) Distribution of the diffONT and pycoMeth predicted regions among the chromosomes. The X chromosome includes more than 95 500 (91 %) of the diffONT predicted regions. (D) Density distribution of the methylation difference of predicted regions by diffONT and pycoMeth . Methylation above 0 means higher methylation in the female sample HG004. Download figure Open in new tab Fig. 9. The diffONT workflow detects potential MSP regions, which differentiate between disease and control samples by containing CpG positions being unmethylated in control and methylated in disease samples (median methylation in disease samples above maximum methylation in control samples). Required input data is one bedmethyl file (e.g. generated by modbam2bed ) per sample, containing information about number of reads and percentage methylation per cytosine for one sample. (A) In a preproccessing step bedmethyl files of different samples are merged and sorted by genomic position into one bedmethyl file. (B) Visualization of CpG methylation information for an exemplary region (exclusively (+) strand data is shown); UCSC CpG islands are annotated in green. (C) The methylation in the control samples is analyzed for each position. If the methylation exceeds the user-defined threshold (default > 10 %) in only one of the control samples, the position is excluded from further analysis, Filter 1. (D, E) Boxplot statistics (0th, 25th, 50th, 75th and 100th percentile) are calculated for disease and control samples of each remaining position. These statistics are provided as tables to the user. (F, G) Only positions where the 50th percentile methylation in disease is above the 100th percentile methylation in control are considered as CpGs of interest. These are screened for potential primer regions. Primer regions should contain at least three differentiating CpGs within (per default) at most 24 nt. While for (F) the distance between the remaining CpGs is too large to form primers, for (G) multiple primers are possible. (H) For all potential primer regions a score is calculated, by summing up the normalized coverage multiplied with the percentage methylation for all remaining cytosine positions within the primer for all disease samples. (I) The table of potential primers and their scores is provided to the user. (J) The best combination of primers is selected based on the highest summed score of two primers within a user-defined region (default 60-400 nt). (K) The score for these MSP regions is determined by summing up the scores of the individual primers. The list containing predicted MSP region scores, GC content, coverage, and MSP region length is provided to the user. Fig. 7 . If many short diffONT MSP regions overlap with one long pycoMeth DMR region, then this results in a high overlap of MSP regions with DMRs and a small number of DMRs overlapping with any MSP result. The number of MSP regions predicted by diffONT for the GIAB dataset is approximately 15 times higher than the number predicted for the ONT dataset, generally indicating more methylation differences between males and females than between healthy and cancerous male cells. Interestingly, the MSP regions predicted by diffONT are distributed unequally among the chromosomes. An extremely high amount of predicted regions can be seen on chromosome X (95 527, ∼91 %), with the second-highest amount of MSP regions predicted on chromosome 1 (1 235), Fig. 8C . This can be explained by the fact, that DNA methylation on the X chromosome reflects sex-specific dosage compensation driven by X-chromosome inactivation (XCI) in the female sample, which has two X chromosomes ( 37 ). Most of the DMRs predicted by pycoMeth are located on chromosome 2 (12 153) and chromosome 1 (11 117) ( Fig. 8C , Tab. 5 ). pycoMeth does not predict an increased amount of DMRs on chromosome X (4 346). Instead, the distribution of DMRs predicted by pycoMeth follows the chromosome length distribution, SFig. S1 . As diffONT is designed to show methylation differences detectable in an MSP, diffONT screens for methylation differences in one direction only, originally higher methylation in tumor samples compared to control samples. pycoMeth on the other hand, is not designed for this specific approach and thus predicts both hyper- and hypomethylated regions. As Fig. 8D displays, all diffONT predicted MSP regions are hypermethylated, i.e., showing higher methylation levels in the female sample. The majority of these predicted MSP regions display a methylation difference exceeding 50 %. As expected, pycoMeth predicts regions with both positive and negative methylation differences between the samples. However, most regions predicted by pycoMeth show a methylation difference below 50 % and thus are difficult to use in an MSP. For comparison of runtime, we excluded basecalling and methylation calling. Running the pycoMeth workflow took 31 590 min. in total. The workflow consists of the short step meth5 create m5 , followed by the methylation-based segmentation of the genome pycometh Meth Seg , which took 31 293 min., and is advised to be performed chromosome-wise. The final step pycometh Meth Comp took 104 min. Despite a longer preprocessing time (421 min.), diffONT is in total much faster than pycoMeth (1216 min.), see Tab. 6 . View this table: View inline View popup Download powerpoint Table 6. Comparison of the main result statistics for diffONT and pycoMeth on the GIAB dataset. Conclusion and Outlook Here we presented diffONT , a tool to predict methylation-specific PCR biomarker regions, based on nanopore methylation sequencing data. diffONT is easily executable via the command line. In this study, we demonstrated that the MSP regions predicted by diffONT form a subset of the DMRs identified by other differential methylation calling tools, such as DSS and metilene . However, this subset has the advantage of showing only distinctly hypermethylated regions compared to the control sample, which is restricted to be nearly unmethylated. While diffONT is slower than metilene , its unique advantage lies in its ability to directly generate regions suitable for primer design in MSP. This feature makes diffONT particularly valuable for clinical applications, such as identifying biomarkers for cancer detection and potentially screening. Designed with a focus on this clinical use, diffONT simplifies and accelerates the development of methylation-specific diagnostic tests for cancer. By combining ease of use with direct applicability, diffONT provides a practical and efficient tool for bridging the gap between epigenomic research and clinical diagnostics. Currently, diffONT does not incorporate allele-specific data ( 13 ) or information regarding the primer annealing temperature. However, this design choice allows users the flexibility to apply their own tools and methods for selecting primers, depending on their specific needs and preferences. Additionally, while diffONT is efficient in identifying MSP regions, its runtime could be further optimized by combining it with other DMR calling tools, such as metilene . Given that metilene includes nearly all regions predicted by diffONT , an effective strategy would be to use metilene for a pre-selection step followed by diffONT for analysis of the pre-selected regions. This combination could significantly reduce runtime while maintaining the accuracy and specificity of the MSP regions identified. These improvements would further strengthen diffONT ’s utility in clinical and research applications. Overall, diffONT presents a unique solution for the clinical use-case of predicting MSP-specific primer regions and thus can help to transfer knowledge of methylation-specific (cancer) markers into clinical diagnostics. Material and Methods The diffONT pipeline diffONT is a Python-based tool available for download on GitHub 2 , designed for analyzing DNA methylation differences between a control group (e.g., healthy individuals) and a disease group (e.g., patients with the same cancer type). It evaluates methylation variations across the entire genome, taking into account sequencing coverage and specifically identifying regions where methylation is absent in the control group but elevated in the disease group. diffONT outputs and scores potential biomarker regions independent of their location relative to known genes, including upstream, downstream, intergenic, and gene body regions. Additionally, it suggests primer sequences suitable for methylation-specific PCR (MSP) to facilitate experimental validation of identified potential biomarkers. Data preprocessing A merged and sorted bedmethyl file, containing data from at least one control and one disease sample (see STab. S1) including an additional column for the sample name, serves as the input for diffONT using the --bedmethyl parameter. The required input bedmethyl file can be generated with the script preprocess.py , which merges multiple bedmethyl files into one bedmethyl file, and sorts it by genomic positions. An additional column defines the origin of the entry for each line. The bedmethyl file contains for each sample for each position and for each strand information about coverage and methylation percentage ( Fig. 9A/B ). Filter 1: Non-methylated CpGs in control samples For the usage in a clinical context, we aim to reduce the number of false positives and require the control samples to show almost no methylation and diseased samples being hypermethylated. Further, we expect that the larger the difference between the two conditions, the more robust the MSP results will be, when using the predicted primers as biomarkers in a clinical screening context. Therefore, we implemented a filter, which removes from the bedmethyl file all positions with a methylation above 10% (default of --minCpGs ) over all reads mapping to this position in at least one control sample, Fig. 9C . This cutoff can be adapted by the user to increase or decrease specificity by setting a lower or higher methylation threshold, respectively, in control samples. Adaption might be preferable if (i) the input data quality changes or (ii) different PCR assays are used, e.g., when using additional TaqMan probing. Calculation of boxplot data Next, for all remaining positions, the methylation values for the disease and control samples are aggregated per position and visualized as boxplots ( Fig. 9D-G ). The 25th and 75th (equals upper and lower quartile, respectively) are calculated with the numpy function nanpercentile , the 50th percentile (equals median) with the function nanmedian . NAN values, where no methylation status could be determined for a sample and position were not taken into account in this way. For the calculation of the 0th and 100th percentile we use the interquartile range iqr = 75th percentile − 25th percentile. The 0th percentile was calculated as the lowest value above 25th percentile −1.5 · iqr . The 100th percentile was calculated as the highest value below 75th percentile +1.5 · iqr . The 0th and 100th percentile represent the highest and lowest value, excluding outliers. Filter 2: Selection for differentiating CpGs We remove all CpGs, which show no strong differences between disease and control sample. Thus, for each position the median disease methylation (50th percentile) has to be above the 100th percentile of the control sample. As a result, for most arbitrary genomic regions, the methylation in disease samples is only moderately elevated, such as in Fig. 9D , and therefore removed, see Fig. 9F . Potential biomarker regions are expected to be hypermethylated over a longer distance in diseased samples, see Fig. 9E and are subjected to the rest of the workflow, see Fig. 9E/G . Filter 3: Selection of MSP primer suitable regions Regions for the MSP primer design are selected based on a defined (i) primer length (default 18-24 nt by --minPrimerLength and --maxPrimerLength ) and a (ii) minimum number of differentially methylated cytosines ( --minCpGs default 3). Thus, the genome is screened for regions which have at least three positions being methylated in the disease samples and unmethylated in the control samples within at most 24 nucleotides. The selected potential MSP regions can contain multiple possible primer positions, see Fig. 9G . Score calculation For all potential primer regions a score is calculated based on the disease sample coverage and disease sample methylation, see Fig. 9H/I . The primer score is calculated as: with N = number of disease samples, M = number of CpGs within the primer, and K = number of CpGs within disease sample i. For all remaining CpGs within a primer, the normalized coverage is multiplied by the methylation percentage for each disease sample. In this way, primers with highly covered and highly methylated cytosines in the disease samples result in a high score. As the control methylation has already been checked in Filter 1 and 2, only the disease sample methylation is taken into account. The coverage is normalized to be comparable between all disease samples to receive relative coverage values for each disease sample. Primer pair formation All potential primer combinations resulting in MSP regions with length between 60 and 400 nt (default, including primers, addressed by --minAmpliconLength and --maxAmpliconLength ) are extracted, with the score being the sum of the two corresponding primer scores, see Fig. 9J/K . Overlapping primer pairs are possible and desired due to a combination of different annealing temperatures and other user-specific requirements. Further output By default, diffONT produces four output files: boxplotData.tsv contains aggregated methylation information per position for all differentially methylated cytosines that meet the criteria of Filter 1 and Filter 2. interestingCpGs.tsv provides per sample, per position coverage and methylation information extracted from the bedmethyl file for all differentially methylated cytosines that fulfill the criteria of Filter 1 and Filter 2. primerData.tsv lists information for each predicted MSP primer. pcrProduct.tsv includes details for every predicted MSP region. If the program is run multiple times, these files will be overwritten unless a different output folder is specified. As further output, the DNA sequence for the primer and MSP regions can be reported, if a genome reference sequence is given as input (with parameter --reference ). For the primers, the DNA sequence is reported in an additional column in the output file primerData.tsv . For the MSP regions the DNA sequences are stored in a fasta file mspRegions.fa . Additionally, annotated genes overlapping the MSP region are reported in an additional column within the MSP region output file pcrProduct.tsv , if an Ensembl ( 38 ) genome annotation is given by the parameter --annotation in gtf or gff file format. This approach is particularly well-suited for analyzing very low coverage data and small sample sizes, as it does not rely on traditional test statistics. While it’s important to acknowledge that lower coverage or fewer replicates may result in a higher likelihood of false positives and false negatives, the method remains robust under these conditions. Additionally, with fewer control samples, the runtime may increase slightly due to the retention of more CpGs for analysis. Often, multiple potential primer regions are located within close proximity. The diffONT github repository https://github.com/rnajena/diffONT/ has the option with a separate script groupPCRproducts.py to merge overlapping regions into large regions. These larger regions can then be used for primer design with other tools (e.g., Thermo Fisher’s Methyl Primer Express ( 39 )). ONT benchmarking dataset for comparison with metilene and DSS The publicly available Oxford Nanopore Technologies (ONT) Benchmark Dataset rrms 2022.07 was accessed in November 2023 from ONT 3 . In the following, we refer to this dataset as “ONT dataset”. The corresponding human reference genome hg38 (GCA 000001405.15) and annotation (GCA 000001405.15) were retrieved from the NCBI ( 40 ) ftp server. The dataset consists of DNA extracted from COLO829 melanoma fibroblasts (ATCC CRL-1974) and COLO829BL normal B lymphoblasts (ATCC CRL-1980) cell-lines, which were established from a 45-year-old male individual with metastatic malignant melanoma. The DNA was sequenced on multiple flow cells per sample, using adaptive sampling to enrich for CpG islands in the human genome, see Tab. 7 . Methylation-specific basecalling has been performed, resulting in bam files containing per CpG and strand orientation methylation information ( 41 ). We calculated the retained coverage per sample using mosdepth (v0.3.3) ( 42 ). We used the technical replicates originating from the different flow cells as biological replicates for the comparison of the different tools. View this table: View inline View popup Download powerpoint Table 7. Benchmarking ONT dataset ( 43 ) consisting of five technical replicates for normal B lymphoblasts (COLO829BL) and melanoma fibroblast (COLO829) cell lines. Each technical replicate was Nanopore sequenced on a different flow cell (FC), resulting in an average coverage of about 7X per replicate. The coverage was calculated using the whole genome coverage calculation tool mosdepth ( 42 ). As input data for diffONT we prepared the benchmark dataset using the modbam2bed package (v0.9.4) ( 44 ). We extracted the methylation information from the ONT output files, which are in bam file format, into bedmethyl-formatted files ( 45 ), see STab. S1. GIAB benchmarking dataset for comparison with pycoMeth We could not use the same benchmarking dataset for pycoMeth as for the other tools because pycoMeth relies on methylation calls from nanopolish ( 9 ). Re-basecalling and methylation calling with nanopolish required the original fast5 raw data, which was corrupted in the ONT benchmarking dataset, rendering it unusable. Therefore, for the comparison of diffONT with pycoMeth , we used the benchmarking dataset from the pycoMeth publication ( 12 ). However, this GIAB dataset does not contain cancer-specific data. As a result, the comparison with metilene and DSS was conducted using the ONT benchmarking dataset, which includes the relevant cancer-specific data. We downloaded the samples HG003 ( 46 ) and HG004 ( 47 ) from the Ashkenazim Trio from the GIAB consortium ( 48 ). These samples are from a healthy human male and female, respectively. Basecalling of the fast5 raw data was performed with Guppy (v6.5.7) 4 , using the model dna r9.4.1 450bps modbases 5mc cg hac.cfg ; methylation calling was performed with nanopolish (v0.14.0) ( 9 ). For MSP region prediction with diffONT , the bam output files generated by Guppy were merged, sorted and indexed with SAMtools (v1.16.1) ( 49 ) and converted into bedmethyl file format using modbam2bed (v0.9.4) 5 , before MSP regions were predicted using diffONT . pycoMeth DMR detection was performed on the GIAB dataset with default parameters, by running first CpG Aggregate , then segmenting the genome into intervals based on methylation by using Meth Seg and finally using Meth Comp of pycoMeth ( 12 ) to check for differential methylation for each interval. Comparison with tools metilene and DSS (DMR) For comparison, the ONT dataset was analyzed with metilene (v0.2-8) ( 11 ), DSS (v2.44.0) ( 14 ), and diffONT . The tool metilene , originally designed for analyzing WGBS and RRBS data, requires DNA methylation information in bedgraph format. An in-house bash script converted the data from bedmethyl into bedgraph file format. In this step, all lines with NAN entries in the bedmethyl file (containing no methylation information) were removed. metilene was run using the default parameter. For runtime comparison of diffONT with metilene , DSS and pycoMeth , all tools were run on a 64 core processor with CPUs of model AMD Opteron(tm) Processor 6376. All tools were run with default settings for number of threads, if not stated otherwise. The overlap of predicted regions reported by diffONT and metilene was determined by using the intersect method from bedtools (v2.31.1) ( 50 ) with the parameter -u and calculating the number of resulting entries. This step was made in both directions, determining the number of metilene DMRs overlapping any region predicted by diffONT and vice versa. DSS is originally developed for the analysis of WGBS and RRBS data. DSS was run in R (v4.2.1) ( 51 ) using the package DSS (v2.44.0) ( 14 ) with default parameters, requiring a minimum length of 50 nt and a minimum number of three CpG sites per DMR, and 50 % of CpG sites in the DMR having p-values below 10 −5 . The overlap of predicted regions on chromosome 17 reported by diffONT , metilene and DSS was determined by using the intersect method from bedtools (v2.31.1) ( 50 ) with the parameter -u and calculating the number of resulting entries. The overlap was calculated in all three directions. Visualization and CpG island annotation The Integrative Genomics Viewer ( igv , version snapshot 05/04/2023) ( 52 ) has been used to visualize the read methylation status for the predicted MSP regions. CpG island annotations for the human genome hg38, which are shown in Fig 9 , Fig. 5 , and SFig. S4C were downloaded from the UCSC genome browser 6 . All plots shown were created using custom-made scripts run under R (v4.2.1) ( 51 ) using the packages ggplot2 (v3.5.1) ( 53 ), ggvenn (v0.1.10) ( 54 ) and dplyr (1.1.4) ( 55 ), unless described otherwise. Declarations Availability of data and materials The code of diffONT is available at github ( https://github.com/rnajena/diffONT/ ) under MIT license. Data used for the comparison as well as an archived version of the github repository is available under Zenodo.org under the DOI 10.5281/zenodo.14501611. Competing interests No competing interest is declared. Funding The work of Daria Meyer was partially financed by oncgnostics GmbH. Additionally, this work was funded by the Deutsche Forschungsgemeinschaft (DFG, German Research Foundation) under Germany’s Excellence Strategy – EXC 2051 – Project-ID 390713860; NFDI4Microbiota (NFDI 28/1); and the Ministry for Economics, Sciences and Digital Society of Thuringia (TMWWDG), under the framework of the Landesprogramm ProDigital (DigLeben-5575/10-9). Authors’ contributions DM, EB, LW and MM have made substantial contributions to the conception and design of the work. All authors read and approved the final manuscript. Supplement View this table: View inline View popup Download powerpoint Table S1. The bedmethyl file format contains summarized methylation information extracted from bam files. The bedMethyl file is a bed9+2 file with the following information in its columns: 1) reference chromosome, 2) Start position in chromosome, 3) End position in chromosome, 4) Name of item, 5) Score from 0-1000. Capped number of reads, 6) Strandedness, plus (+), minus (-), or unknown (.), 7) Start of where display should be thick (start codon) – data not shown, 8) End of where display should be thick (stop codon) – data not shown, 9) Color value (RGB), 10) Coverage, or number of reads, 11) Percentage of reads that show methylation at this position in the genome. Download figure Open in new tab Fig. S1. (A) Histogram of the coverage distribution for all regions predicted by diffONT on the ONT dataset. Data was binned for coverage, mean coverage per region of control samples and tumor samples is shown in orange and green, respectively. (B) Barplot showing the size of the different chromosomes, based on the hg38 human reference genome. Chromosome length decreases from chromosome 1 to chromosome 22. The size of chromosome X is increased compared to chromosome Y. View this table: View inline View popup Download powerpoint Table S2. Sequencing coverage per chromosome per flowcell, calculated with mosdepth ( 42 ). Sequencing coverage shows differences between the chromosomes over all samples. Especially chromosome Y but also chromosome X show a low coverage (Coverage below 1X for all COLO829 flowcells and below 2X for all COLO829BL samples for chromosome Y). On the other hand, chromosomes 19 and 22 show the highest overall coverage with a mean coverage of 16.76 X on chromosome 19 and 12.89 X on chromosome 22. Download figure Open in new tab Fig. S2. Histogram of the absolute number of (A) Genes, (B) CpGs and (C) CpG islands among the hg38 human genome per chromosome, and (D)-(F) the corresponding densities in relation to chromosome length in [Mb]. Data is based on the human genome version hg38 (GCA 000001405.15). View this table: View inline View popup Download powerpoint Table S3. Result statistics for the ten highest scored MSP regions predicted by diffONT . Download figure Open in new tab Fig. S3. (A) Boxplot showing the ratio of CpGs of interest to total CpGs per chromosome. The fraction was calculated window-wise (size 10 000 nt), which were shifted by half the window length through the genome. The highest median fraction of CpGs of interest is observed on chromosome 4. (B) Boxplot showing the ratio of CpGs of interest to total CpGs within the entire human genome, by using a sliding window approach. CpGs of interest are those CpGs which are passing filter 1 and filter 2. The ratio was calculated for windows of the sizes 1 000, 10 000, 100 000 and 1 000 000 nt. Windows were shifted by half the respective window length through the genome. Larger window sizes result in higher median proportion of CpGs of interest. Download figure Open in new tab Fig. S4. (A) Histogram of the methylation difference values of the regions predicted by diffONT on the ONT dataset. The methylation difference is calculated by subtracting the average melanoma fibroblast samples’ methylation from the average normal B lymphoblast samples’ methylation. (B) Visualization of the MSP region score and average methylation in tumor samples for all MSP regions, predicted by diffONT on the ONT dataset. The coverage per region shows a gap between 5-8X coverage for the diseased samples. (C) Visualization of one of the top 50 reported MSP regions predicted by diffONT on the ONT dataset using the Integrative Genomics Viewer (IGV). At the top the region predicted by diffONT is annotated in blue; at the bottom the gene TCF20 is annotated in blue, and an annotated CpG island in green. The 5mC methylation status is color-coded, red = methylated, blue = unmethylated. The melanoma fibroblast sample is shown on top (5x), the normal B lymphoblast sample at the bottom (5x). The predicted MSP region shows strong methylation differences overlapping the TCF20 5’ UTR and promoter region and CpG island. Acknowledgments The authors thank Alfred Hansel and Martina Schmitz for helpful comments and feedback throughout the project. Footnotes ↵ 1 https://github.com/nanoporetech/dorado ↵ 2 https://github.com/rnajena/diffONT/ ↵ 3 https://registry.opendata.aws/ont-open-data , accessed via aws s3 ls --no-sign-request, s3://ont-open-data/rrms_2022.07/ ↵ 4 Nanopore Community. https://nanoporetech.com/community . ↵ 5 https://github.com/epi2me-labs/modbam2bed ↵ 6 https://genome.ucsc.edu/cgi-bin/hgTables , accessed in June 2021 References 1. ↵ D. Zilberman and S. Henikoff , “Genome-wide analysis of dna methylation patterns,” 2007 . 2. ↵ M. Frommer , L. E. McDonald , D. S. Millar , C. M. Collis , F. Watt , G. W. Grigg , P. L. Molloy , and C. L. Paul , “ A genomic sequencing protocol that yields a positive display of 5-methylcytosine residues in individual dna strands .,” Proceedings of the National Academy of Sciences , vol. 89 , no. 5 , pp. 1827 – 1831 , 1992 . OpenUrl Abstract / FREE Full Text 3. ↵ P. W. Laird , “ The power and the promise of dna methylation markers ,” Nature Reviews Cancer , vol. 3 , no. 4 , pp. 253 – 266 , 2003 . OpenUrl CrossRef PubMed Web of Science 4. ↵ W. J. Locke , D. Guanzon , C. Ma , Y. J. Liew , K. R. Duesing , K. Y. Fung , and J. P. Ross , “ Dna methylation cancer biomarkers: translation to the clinic ,” Frontiers in genetics , vol. 10 , p. 1150 , 2019 . OpenUrl CrossRef PubMed 5. ↵ J. G. Herman , J. R. Graff , S. Myöhänen , B. D. Nelkin , and S. B. Baylin , “ Methylation-specific pcr: a novel pcr assay for methylation status of cpg islands .,” Proceedings of the national academy of sciences , vol. 93 , no. 18 , pp. 9821 – 9826 , 1996 . OpenUrl Abstract / FREE Full Text 6. ↵ C. Dieffenbach , T. Lowe , G. Dveksler , et al. , “ General concepts for pcr primer design ,” PCR methods appl , vol. 3 , no. 3 , pp. S30 – S37 , 1993 . OpenUrl CrossRef PubMed Web of Science 7. ↵ L.-C. Li , “ Designing pcr primer for dna methylation mapping ,” PCR Primer Design , pp. 370 – 383 , 2007 . 8. ↵ V. K. Singh and A. Kumar , “ Pcr primer design ,” Mol. Biol. Today , vol. 2 , no. 2 , pp. 27 – 32 , 2001 . OpenUrl 9. ↵ J. T. Simpson , R. E. Workman , P. Zuzarte , M. David , L. Dursi , and W. Timp , “ Detecting dna cytosine methylation using nanopore sequencing ,” Nature methods , vol. 14 , no. 4 , pp. 407 – 410 , 2017 . OpenUrl CrossRef PubMed 10. ↵ Y. Park and H. Wu , “ Differential methylation analysis for bs-seq data under general experimental design ,” Bioinformatics , vol. 32 , no. 10 , pp. 1446 – 1453 , 2016 . OpenUrl CrossRef PubMed 11. ↵ F. Jühling , H. Kretzmer , S. H. Bernhart , C. Otto , P. F. Stadler , and S. Hoffmann , “ Metilene: fast and sensitive calling of differentially methylated regions from bisulfite sequencing data ,” Genome research , vol. 26 , no. 2 , pp. 256 – 262 , 2016 . OpenUrl Abstract / FREE Full Text 12. ↵ R. Snajder , A. Leger , O. Stegle , and M. J. Bonder , “ pycometh: a toolbox for differential methylation testing from nanopore methylation calls ,” Genome Biology , vol. 24 , no. 1 , pp. 1 – 19 , 2023 . OpenUrl CrossRef PubMed 13. ↵ V. Akbari , J.-M. Garant , K. O’Neill , P. Pandoh , R. Moore , M. A. Marra , M. Hirst , and S. J. Jones , “ Megabasescale methylation phasing using nanopore long reads and nanomethphase ,” Genome biology , vol. 22 , pp. 1 – 21 , 2021 . OpenUrl CrossRef PubMed 14. ↵ H. Feng , K. N. Conneely , and H. Wu , “ A bayesian hierarchical model to detect differentially methylated loci from single nucleotide resolution sequencing data ,” Nucleic acids research , vol. 42 , no. 8 , pp. e69 – e69 , 2014 . OpenUrl CrossRef PubMed 15. ↵ B. Gong and E. Purdom , “ Methcp: differentially methylated region detection with change point models ,” Journal of Computational Biology , vol. 27 , no. 4 , pp. 458 – 471 , 2020 . OpenUrl CrossRef PubMed 16. ↵ Y. Piao , W. Xu , K. H. Park , K. H. Ryu , and R. Xiang , “ Comprehensive evaluation of differential methylation analysis methods for bisulfite sequencing data ,” International Journal of Environmental Research and Public Health , vol. 18 , no. 15 , p. 7975 , 2021 . OpenUrl CrossRef 17. ↵ D. E. Condon , P. V. Tran , Y.-C. Lien , J. Schug , K. Georgieff , R. A. Simmons , and K.-J. Won , “ Defiant:(dmrs: easy, fast, identification and annotation) identifies differentially methylated regions from irondeficient rat hippocampus ,” BMC bioinformatics , vol. 19 , pp. 1 – 12 , 2018 . OpenUrl CrossRef PubMed 18. ↵ H. Fujii , M. A. Biel , W. Zhou , S. A. Weitzman , S. B. Baylin , and E. Gabrielson , “ Methylation of the hic-1 candidate tumor suppressor gene in human breast cancer ,” Oncogene , vol. 16 , no. 16 , pp. 2159 – 2164 , 1998 . OpenUrl CrossRef PubMed Web of Science 19. ↵ C. Vicente , A. Conchillo , M.A. García-Sánchez , and M. D. Odero , “ The role of the gata2 transcription factor in normal and malignant hematopoiesis ,” Critical reviews in oncology/hematology , vol. 82 , no. 1 , pp. 1 – 17 , 2012 . OpenUrl CrossRef PubMed 20. ↵ J. Tian and Y. Wang , “ The lncrnas mir600hg and tspoap1as1 may potentially act as biomarkers for predicting pancreatic cancer ,” Translational Cancer Research , vol. 9 , no. 2 , p. 809 , 2020 . OpenUrl CrossRef PubMed 21. ↵ E. I. Velazquez-Villarreal , S. Maheshwari , J. Sorenson , I. T. Fiddes , V. Kumar , Y. Yin , M. G. Webb , C. Catalanotti , M. Grigorova , P. A. Edwards , et al. , “ Single-cell sequencing of genomic dna resolves sub-clonal heterogeneity in a melanoma cell line ,” Communications biology , vol. 3 , no. 1 , p. 318 , 2020 . OpenUrl CrossRef PubMed 22. ↵ S. Galiegue , O. Jbilo , T. Combes , E. Bribes , P. Carayon , G. Le Fur , and P. Casellas , “ Cloning and characterization of prax-1: a new protein that specifically interacts with the peripheral benzodiazepine receptor ,” Journal of Biological Chemistry , vol. 274 , no. 5 , pp. 2938 – 2952 , 1999 . OpenUrl Abstract / FREE Full Text 23. ↵ A. M. Deaton and A. Bird , “ Cpg islands and the regulation of transcription ,” Genes & development , vol. 25 , no. 10 , pp. 1010 – 1022 , 2011 . OpenUrl Abstract / FREE Full Text 24. ↵ P. J. Balwierz , P. Carninci , C. O. Daub , J. Kawai , Y. Hayashizaki , W. Van Belle , C. Beisel , and E. van Nimwegen , “ Methods for analyzing deep sequencing expression data: constructing the human and mouse promoterome with deepcage data ,” Genome biology , vol. 10 , pp. 1 – 21 , 2009 . OpenUrl CrossRef 25. ↵ F. Larsen , G. Gundersen , R. Lopez , and H. Prydz , “ Cpg islands as gene markers in the human genome ,” Genomics , vol. 13 , no. 4 , pp. 1095 – 1107 , 1992 . OpenUrl CrossRef PubMed Web of Science 26. ↵ L. Ponger , L. Duret , and D. Mouchiroud , “ Determinants of cpg islands: expression in early embryo and isochore structure ,” Genome research , vol. 11 , no. 11 , pp. 1854 – 1860 , 2001 . OpenUrl Abstract / FREE Full Text 27. ↵ P. Hevezi , F. Vences-Catalan , J. L. Maravillas-Montero , C. A. White , P. Casali , L. Llorente , J. Jakez-Ocampo , G. Lima , N. Vilches-Cisneros , J.P. Flores-Gutiérrez , et al. , “ Tspan33 is a novel marker of activated and malignant b cells ,” Clinical Immunology , vol. 149 , no. 3 , pp. 388 – 399 , 2013 . OpenUrl CrossRef PubMed 28. ↵ E. Torti , B. Keren , E. E. Palmer , Z. Zhu , A. Afenjar , I. J. Anderson , M. V. Andrews , C. Atkinson , M. Au , S. A. Berry , et al. , “ Variants in tcf20 in neurodevelopmental disability: description of 27 new patients and review of literature ,” Genetics in Medicine , vol. 21 , no. 9 , pp. 2036 – 2042 , 2019 . OpenUrl CrossRef PubMed 29. ↵ M. Taniwaki , Y. Daigo , N. Ishikawa , A. Takano , T. Tsunoda , W. Yasui , K. Inai , N. Kohno , and Y. Nakamura , “ Gene expression profiles of small-cell lung cancers: molecular signatures of lung cancer ,” International journal of oncology , vol. 29 , no. 3 , pp. 567 – 575 , 2006 . OpenUrl PubMed Web of Science 30. ↵ M. Bacac , E. Migliavacca , J.-C. Stehle , T. McKee , M. Delorenzi , J.-M. Coindre , L. Guillou , and I. Stamenkovic , “ A gene expression signature that distinguishes desmoid tumours from nodular fasciitis ,” The Journal of Pathology: A Journal of the Pathological Society of Great Britain and Ireland , vol. 208 , no. 4 , pp. 543 – 553 , 2006 . OpenUrl 31. ↵ M.-E. López , D. Roquis , C. Becker , B. Denoyes , and E. Bucher , “ Dna methylation dynamics during stress response in woodland strawberry (fragaria vesca) ,” Horticulture research , vol. 9 , p. uhac174 , 2022 . OpenUrl CrossRef 32. ↵ R. Gilsbach , M. Schwaderer , S. Preissl , B. A. Grüning , D. Kranzhöfer , P. Schneider , T. G. Nührenberg , S. Mulero-Navarro , D. Weichenhan , C. Braun , et al. , “ Distinct epigenetic programs regulate cardiac myocyte development and disease in the human heart in vivo ,” Nature communications , vol. 9 , no. 1 , p. 391 , 2018 . OpenUrl CrossRef PubMed 33. ↵ N. Hamazaki , H. Kyogoku , H. Araki , F. Miura , C. Horikawa , N. Hamada , S. Shimamoto , O. Hikabe , K. Nakashima , T. S. Kitajima , et al. , “ Reconstitution of the oocyte transcriptional network with transcription factors ,” Nature , vol. 589 , no. 7841 , pp. 264 – 269 , 2021 . OpenUrl CrossRef PubMed 34. ↵ X. Cui , Q. Lin , M. Chen , Y. Wang , Y. Wang , Y. Wang , J. Tao , H. Yin , and T. Zhao , “ Long-read sequencing unveils novel somatic variants and methylation patterns in the genetic information system of early lung cancer ,” Computers in Biology and Medicine , vol. 171 , p. 108174 , 2024 . OpenUrl CrossRef PubMed 35. ↵ S. Gigante , Q. Gouil , A. Lucattini , A. Keniry , T. Beck , M. Tinning , L. Gordon , C. Woodruff , T. P. Speed , M. E. Blewitt , et al. , “ Using long-read sequencing to detect imprinted dna methylation ,” Nucleic acids research , vol. 47 , no. 8 , pp. e46 – e46 , 2019 . OpenUrl CrossRef PubMed 36. ↵ J. Wang , C. Zhang , L. Zhang , H.-j. Yao , X. Liu , Y. Shi , J. Zhao , X. Bo , H. Chen , and L. Li , “ Comparative study on genomic and epigenomic profiles of retinoblastoma or tuberous sclerosis complex via nanopore sequencing and a joint screening framework ,” Cancer Gene Therapy , vol. 31 , no. 3 , pp. 439 – 453 , 2024 . OpenUrl CrossRef PubMed 37. ↵ C. G. Duncan , S. A. Grimm , D. L. Morgan , P. R. Bushel , B. D. Bennett , J. D. Roberts , F. L. Tyson , B. A. Merrick , and P. A. Wade , “ Dosage compensation and dna methylation landscape of the x chromosome in mouse liver ,” Scientific reports , vol. 8 , no. 1 , p. 10138 , 2018 . OpenUrl CrossRef PubMed 38. ↵ B. L. Aken , S. Ayling , D. Barrell , L. Clarke , V. Curwen , S. Fairley , J. Fernandez Banet , K. Billis , C. García Girón , T. Hourlier , et al. , “ The ensembl gene annotation system ,” Database , vol. 2016 , p. baw093 , 2016 . OpenUrl CrossRef PubMed 39. ↵ “Methyl primer express™.” https://www.thermofisher.com/order/catalog/product/de/de/4376041 . Accessed: May 2024 . 40. ↵ D. L. Wheeler , T. Barrett , D. A. Benson , S. H. Bryant , K. Canese , V. Chetvernin , D. M. Church , M. DiCuccio , R. Edgar , S. Federhen , et al. , “ Database resources of the national center for biotechnology information ,” Nucleic acids research , vol. 36 , no. Suppl 1 , pp. D13 – D21 , 2007 . OpenUrl CrossRef PubMed Web of Science 41. ↵ “Oxford nanopore technologies. rrms.” https://community . nanoporetech.com/attachments/7600/download. Accessed: June 2024 . 42. ↵ B. S. Pedersen and A. R. Quinlan , “ Mosdepth: quick coverage calculation for genomes and exomes ,” Bioinformatics , vol. 34 , no. 5 , pp. 867 – 868 , 2018 . OpenUrl CrossRef PubMed 43. ↵ “Oxford nanopore technologies benchmark dataset rrms 2022.07.” https://registry.opendata.aws/ont-open-data/rrms_2022.07 . Accessed: November 2023 . 44. ↵ “Oxford nanopore technologies. modbam2bed.” https://github.com/epi2me-labs/modbam2bed . Accessed: October 2023 . 45. ↵ “encodeproject.” https://www.encodeproject.org/data-standards/wgbs/ . Accessed: October 2023 . 46. ↵ “Giab hg003 fast5 data.” https://s3-us-west-2.amazonaws . com/human-pangenomics/NHGRI_UCSC_panel/HG003/nanopore/GM24149_1.fast5.tar.gz. Accessed: August 2024 . 47. ↵ “Giab hg004 fast5 data.” https://s3-us-west-2.amazonaws . com/human-pangenomics/NHGRI_UCSC_panel/HG004/nanopore/GM24143_1.fast5.tar.gz. Accessed: August 2024 . 48. ↵ J. M. Zook , D. Catoe , J. McDaniel , L. Vang , N. Spies , A. Sidow , Z. Weng , Y. Liu , C. E. Mason , N. Alexander , et al. , “ Extensive sequencing of seven human genomes to characterize benchmark reference materials ,” Scientific data , vol. 3 , no. 1 , pp. 1 – 26 , 2016 . OpenUrl 49. ↵ P. Danecek , J. K. Bonfield , J. Liddle , J. Marshall , V. Ohan , M. O. Pollard , A. Whitwham , T. Keane , S. A. McCarthy , R. M. Davies , et al. , “ Twelve years of samtools and bcftools ,” Gigascience , vol. 10 , no. 2 , p. giab008 , 2021 . OpenUrl CrossRef PubMed 50. ↵ A. R. Quinlan and I. M. Hall , “ Bedtools: a flexible suite of utilities for comparing genomic features ,” Bioinformatics , vol. 26 , no. 6 , pp. 841 – 842 , 2010 . OpenUrl CrossRef PubMed Web of Science 51. ↵ R Core Team , R: A Language and Environment for Statistical Computing . R Foundation for Statistical Computing , Vienna, Austria , 2022 . 52. ↵ J. T. Robinson , H. Thorvaldsdóttir , W. Winckler , M. Guttman , E. S. Lander , G. Getz , and J. P. Mesirov , “ Integrative genomics viewer ,” Nature biotechnology , vol. 29 , no. 1 , pp. 24 – 26 , 2011 . OpenUrl CrossRef PubMed Web of Science 53. ↵ H. Wickham , ggplot2: Elegant Graphics for Data Analysis . Springer-Verlag New York , 2016 . 54. ↵ L. Yan , ggvenn: Draw Venn Diagram by ‘ggplot2’ , 2023 . R package version 0.1.10 . 55. ↵ H. Wickham , R. François , L. Henry , K. Müller , and D. Vaughan , dplyr: A Grammar of Data Manipulation , 2023 . R package version 1.1.4 . View the discussion thread. Back to top Previous Next Posted February 20, 2025. Download PDF Email Thank you for your interest in spreading the word about bioRxiv. NOTE: Your email address is requested solely to identify you as the sender of this article. Your Email * Your Name * Send To * Enter multiple addresses on separate lines or separate them with commas. You are going to email the following diffONT: predicting methylation-specific PCR biomarkers based on nanopore sequencing data for clinical application Message Subject (Your Name) has forwarded a page to you from bioRxiv Message Body (Your Name) thought you would like to see this page from the bioRxiv website. Your Personal Message CAPTCHA This question is for testing whether or not you are a human visitor and to prevent automated spam submissions. Share diffONT: predicting methylation-specific PCR biomarkers based on nanopore sequencing data for clinical application Daria Meyer , Emanuel Barth , Laura Wiehle , Manja Marz bioRxiv 2025.02.17.638597; doi: https://doi.org/10.1101/2025.02.17.638597 Share This Article: Copy Citation Tools diffONT: predicting methylation-specific PCR biomarkers based on nanopore sequencing data for clinical application Daria Meyer , Emanuel Barth , Laura Wiehle , Manja Marz bioRxiv 2025.02.17.638597; doi: https://doi.org/10.1101/2025.02.17.638597 Citation Manager Formats BibTeX Bookends EasyBib EndNote (tagged) EndNote 8 (xml) Medlars Mendeley Papers RefWorks Tagged Ref Manager RIS Zotero Tweet Widget Facebook Like Google Plus One Subject Area Bioinformatics Subject Areas All Articles Animal Behavior and Cognition (7635) Biochemistry (17691) Bioengineering (13892) Bioinformatics (41937) Biophysics (21452) Cancer Biology (18588) Cell Biology (25504) Clinical Trials (138) Developmental Biology (13378) Ecology (19899) Epidemiology (2067) Evolutionary Biology (24320) Genetics (15609) Genomics (22506) Immunology (17736) Microbiology (40394) Molecular Biology (17181) Neuroscience (88605) Paleontology (666) Pathology (2832) Pharmacology and Toxicology (4824) Physiology (7641) Plant Biology (15156) Scientific Communication and Education (2045) Synthetic Biology (4294) Systems Biology (9825) Zoology (2271)