Application of the Nicking Loop™ targeted library preparation method to DNBSEQ™ sequencing

Circular NGS platform, DNBSEQ™, Nicking Loop, RCA 15 preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for thisthis version posted March 12, 2026. ; https://doi.org/10.64898/2026.03.10.710732doi: bioRxiv preprint 2 FUNDING 16 This study was funded by venture capital from Voima Ventures (Helsinki, Finland), 17 Avohoidon Tutkimussaatio (Espoo, Finland), Business Finland (Helsinki, Finland) and 18 Almaral (Kaarina, Finland). 19 CONFLICT OF INTEREST 20 M.T. is the Chief Executive Officer of Genomill Health Inc. J.P.P is the Chief Technology 21 Officer of Genomill Health Inc. J.L. is a medical advisor at Genomill. J.B., J.L., M.T. hold 22 equity in Genomill. S.A., N.L., A.K., T.H., A.M., T.R., J.K., J.B., J.L., M.T., J.P.P. are 23 entitled to stock options in Genomill. The authors S.A., N.L., A.K., T.H., A.M., T.R., J.K. 24 and J.P.P. are currently employed at Genomill. R.H. is a senior scientist at Complete 25 Genomics Inc. C.X. is the vice president of Complete Genomics Inc. Both R.H. and C.X. do 26 not hold stock or stock option in Complete Genomics Inc. The research detailed in this 27 manuscript is related to Genomill patents EP4332238, TWI868882 and US11970736. 28 Genomill is in the process of filing patents related to this work, and S.A., N.L., A.K., T.H., 29 A.M., T.R., J.K., M.T., J.P.P. are named inventors on these patent applications. J.B. has 30 received honoraria from Novo Nordisk and Boehringer Ingelheim. In preparation of this 31 manuscript, every effort has been made to ensure that the research is conducted and presented 32 objectively. The design of the study, data collection, analysis, interpretation, and the writing 33 of the manuscript were conducted independently of Genomill's commercial interests. The 34 authors affirm that the information provided here is accurate and complete to the best of their 35 knowledge. 36 preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for thisthis version posted March 12, 2026. ; https://doi.org/10.64898/2026.03.10.710732doi: bioRxiv preprint 3

37 Nicking Loop™ is a PCR-free targeted library preparation method that combines conversion 38 of linear target DNA into circular single-stranded DNA (CssDNA) library with early sample 39 indexing in a single step. The resulting CssDNA libraries can be either directly sequenced or 40 optionally amplified, offering maximum flexibility across sequencing applications. 41 This study demonstrates the compatibility of Nicking Loop™ circular libraries with a MGI’s 42 DNBSEQ™ platform. Compatibility was evaluated against established linear Nicking 43 Loop™ libraries sequenced on Illumina MiSeq platform. Using synthetic reference samples 44 with defined variant allele frequencies, Nicking Loop™ method demonstrated matching 45 performance across both library formats and sequencing platforms. Key quality metrics, 46 including unique molecular identifier (UMI) distributions, error profiles and VAF detection, 47 were all highly consistent. Both library types generated over 97% singleton UMIs, indicating 48 uniform template sampling, and VAF measurements were strongly concordant across 49 platforms (Spearman’s ρ = 0.939). Collectively, these findings demonstrate that Nicking 50 Loop™ method is directly applicable to circular NGS platforms, such as DNBSEQ™, 51 strongly supporting its use as a platform-agnostic library preparation strategy for targeted 52 sequencing applications. 53 preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for thisthis version posted March 12, 2026. ; https://doi.org/10.64898/2026.03.10.710732doi: bioRxiv preprint 4 ABBREVIATIONS 54 CssDNA Circular single-stranded DNA 55 DNB DNA nanoball 56 dsDNA Double-stranded DNA 57 NGS Next-generation sequencing 58 ssDNA Single-stranded DNA 59 RCA Rolling-circle amplification 60 RT Room temperature 61 UMI Unique molecular identifier 62 VAF Variant allele frequency 63 preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for thisthis version posted March 12, 2026. ; https://doi.org/10.64898/2026.03.10.710732doi: bioRxiv preprint 5

64 The field of next-generation sequencing (NGS) expanded rapidly over the past decade. In 65 addition to established short-read sequencing platforms such as Illumina, technologies 66 developed by MGI, Element Biosciences, and Pacific Biosciences introduced alternative 67 sequencing chemistries, including the use of circular templates and long-read sequencing 68 strategies1–3. However, many library preparation workflows are optimized for linear, 69 Illumina-like libraries and often require additional conversion steps to generate libraries 70 compatible with circular approaches4. In addition, library preparation protocols for newer 71 NGS platforms are commonly optimized for whole-genome sequencing applications 72 primarily used in research 5–7. In contrast, targeted sequencing approaches remain widely 73 used in clinical and diagnostic settings due to cost-efficiency, high-depth coverage, and assay 74 design considerations8,9. 75 Nicking Loop™ is a targeted library preparation method that converts linear DNA into 76 circular single-stranded DNA (CssDNA) while introducing sample-specific index at the early 77 step of the workflow10. Early sample indexing enables direct sequencing of circular libraries 78 following conversion or pooling of samples prior to further library amplification if desired. 79 Nicking Loop™ leverages linear amplification to limit error propagation, which is a key 80 attribute of circular NGS platforms. These features make Nicking Loop™ well suited for 81 sequencing platforms that utilize circular library formats. This compatibility is particularly 82 relevant for applications where detection of low-frequency variants is critical, including 83 detection of somatic variants from circulating cell-free DNA or tissue, and infectious disease 84 testing. 85 The Nicking Loop™ has been previously evaluated on multiple sequencing platforms, and 86 this study extends the evaluation to MGI’s DNBSEQ™ platform. Here, the performance of 87 circular Nicking Loop™ libraries sequenced on DNBSEQ™-G99 is compared to linear 88 preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for thisthis version posted March 12, 2026. ; https://doi.org/10.64898/2026.03.10.710732doi: bioRxiv preprint 6 Nicking Loop™ libraries sequenced on Illumina MiSeq. Platform compatibility is assessed 89 using key performance metrics, including unique molecular identifier distributions, error 90 profiles and variant allele frequency (VAF) profiling capabilities, with early sample indices 91 used for read assignment on DNBSEQ™ platform. 92

93 Oligonucleotides 94 The components of a Nicking Loop construct: the Loop carrying the sample index, left and 95 right probes, and the bridge oligonucleotide and two synthetic DNA pools used as target 96 DNA were produced by IDT (Coralville, IA). Left and right probes were pre-annealed with 97 the bridge oligonucleotide to form a probe construct. The DNA pools contained ten 98 templates. Individual templates were distinguished by three pool-specific complementary 99 nucleotides. The pools were mixed in different ratios to mimic the VAF of 0%, 1%, 5%, 100 10%, and 20%. 101 Preparation of Nicking Loop™ converted CssDNA and RCA amplification 102 To prepare Nicking Loop™ converted CssDNA from linear target DNA, 2.5 fmol of linear 103 synthetic target pools were combined with 10 fmol of probe constructs (probes 6 & 7 were 104 loaded at 1 fmol instead) and 200 fmol of Loop oligomer in 1× Ampligase buffer (LGC 105 Biosearch Technologies, Hoddesdon, UK). To induce hybridization, the mixture was 106 denatured at 95°C for 5 minutes, cooled stepwise to 55°C (10°C steps with 2.5 minutes hold), 107 and incubated at 55°C for 2 hours. To circularize the hybridization products, the reactions 108 were supplemented with: 0.2 U Phusion High-Fidelity DNA Polymerase (Thermo Fisher 109 Scientific, Waltham, MA), 0.2 U Ampligase (LGC Biosearch Technologies), and 10 mM 110 dNTPs (Thermo Fisher Scientific) in 1 × Ampligase Reaction buffer (LGC Biosearch 111 Technologies) and incubated at 55°C for 40 minutes. 112 preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for thisthis version posted March 12, 2026. ; https://doi.org/10.64898/2026.03.10.710732doi: bioRxiv preprint 7 To degrade residual linear DNA, 1 µL of each Thermolabile Exonuclease I and RecJf (NEB, 113 Ipswich, MA) was added to the reaction, and incubated at 37°C for 45 minutes, followed by 114 heat inactivation at 80°C for 10 minutes. Nicking Loop™ converted CssDNA templates were 115 purified using a 2× volume of AMPure XP beads (Beckman Coulter, Brea, CA). 116 Nicking Loop™ converted CssDNA templates prepared by conversion of linear DNA were 117 amplified by rolling-circle amplification (RCA). In the first step, 10 µL of purified CssDNA 118 templates were pre-annealed with 12.5 pmol of universal amplification primer in 1× 119 EquiPhi29 buffer (Thermo Fisher Scientific). The mixture was denatured at 95°C for 5 120 minutes and gradually cooled down to room temperature (RT) for 20 minutes. To produce 121 RCA concatemers, the reaction was supplemented with 10 mM dNTPs, 100 mM DTT, and 122 0.5 U of EquiPhi29 (Thermo Fisher Scientific) followed by incubation at 45°C for 90 123 minutes, and subsequent heat inactivation at 95°C for 10 minutes. 124 Circular library preparation for DNBSEQ™-99 125 The circular Nicking Loop™ libraries were prepared from RCA-generated concatemers. To 126 promote folding of the concatemers, reactions were supplemented with nuclease-free water 127 and 1× rCutSmart buffer (NEB), incubated at 95°C for 5 minutes, and gradually cooled to 128 RT. Folded concatemers formed a hairpin structure with a dsDNA restriction recognition site 129 for the nicking enzyme. To produce linear monomers, the recognition site was digested with 130 10 U of Nb.BsrDI (NEB) or 15 U of Nb.BbvCI (NEB) at 37°C for 90 minutes, followed by 131 heat inactivation at 80°C for 10 minutes. The resulting linear monomers were purified using 132 1.8× volume of AMPure XP beads (Beckman Coulter). 133 To initiate folding of monomers into a circular Nicking Loop structure with a nick, 10 µL of 134 purified linear monomers was mixed with 7 µL of nuclease-free water, denatured at 95°C for 135 5 minutes, and gradually cooled to RT. The nick was subsequently ligated by addition of 1 136 µL of T4 DNA ligase and 2 µL of 10× T4 DNA ligase buffer (NEB), followed by incubation 137 preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for thisthis version posted March 12, 2026. ; https://doi.org/10.64898/2026.03.10.710732doi: bioRxiv preprint 8 at 32°C for 30 minutes. Any residual linear ssDNA was degraded by 1 µL each of 138 Thermolabile Exonuclease I and RecJf (NEB). The reactions were incubated at 37°C for 45 139 minutes, then heat-inactivated at 95°C for 10 minutes. The final circular library was purified 140 using a 2× volume of AMPure XP beads (Beckman Coulter). The libraries were quantified 141 using the Qubit™ dsDNA HS Assay Kit and Qubit™ 4 Fluorometer (Thermo Fisher 142 Scientific). A schematic of the circular Nicking Loop™ library is provided in Supplementary 143 Figure. 1. 144 DNBSEQ™-G99 Sequencing 145 Nicking Loop™ libraries were amplified into DNA nanoballs (DNBs) and sequenced on the 146 DNBSEQ™-G99 platform following the manufacturer’s system guide (Complete Genomics, 147 San Jose, CA). Briefly, for DNB preparation, 250 fmol of pooled library was combined with 148 10 μL of Make DNB Buffer containing primer for RCA amplification. Then appropriate 149 amount of TE buffer was added to make the final volume of 20 μL for primer hybridization. 150 Hybridization was performed with sequential incubations at 95 °C, 65 °C, and 40 °C for 1 151 minute each. Subsequently, 20 μL of Make DNB Enzyme Mix I and 2 μL of Make DNB 152 Enzyme Mix II (LC) were added, and DNB amplification was carried out at 30 °C for 30 153 minutes. The reaction was terminated by adding 10 μL of Stop DNB Reaction Buffer. DNB 154 concentration was quantified using the Qubit ssDNA Assay Kit (Thermo Fisher Scientific, 155 PN #Q10212) by using 2 μL of above DNB solution. Then, 21 μL of DNB solution was 156 mixed with 7 μL of DNB Load Buffer II and 1 μL of Make DNB Enzyme Mix II (LC), then 157 loaded onto the flow cell according to manufacturer’s system guide. Sequencing was 158 performed using the DNBSEQ™-G99RS High-Throughput Sequencing Set (PN #940-159 001717-00, Model APP-D FCL PE300, Version 1.0). Run parameters were configured for 160 paired-end sequencing with Read 1 of 66 bp, Read 2 of 121 bp, and a 45 bp barcode read. 161 Custom-designed Make DNB primers, sequencing primers, and MDA primers based on the 162 preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for thisthis version posted March 12, 2026. ; https://doi.org/10.64898/2026.03.10.710732doi: bioRxiv preprint 9 Nicking Loop™ library structure were used for DNB amplification and library sequencing 163 (Supplementary Fig. 1). 164 Library preparation and sequencing with Illumina MiSeq 165 The linear Nicking Loop™ libraries were produced from RCA-generated concatemers (see 166 section Preparation of Nicking Loop™ converted CssDNA and RCA amplification). To 167 introduce flowcell binding sequences, 2 µL of RCA concatemers was subjected to 168 minimal-cycle PCR using 10 µM dual-indexed primers with Phusion HS II DNA Polymerase 169 (Thermo Fisher Scientific), following the manufacturer’s guidelines. PCR was performed 170 with a combined annealing and extension step at 72°C for 14 cycles. All PCR products were 171 gel-purified using the Monarch DNA Gel Extraction kit (NEB). The linear libraries were 172 quantified using the Qubit™ dsDNA HS Assay Kit and the Qubit™ 4 Fluorometer (Thermo 173 Fisher Scientific). Sequencing was performed on the Illumina MiSeq platform, using the 174 MiSeq Reagent Kit v3 for a 2 × 300 bp paired-end run (Illumina). The schematic of the linear 175 Nicking Loop™ library structure is provided in Supplementary Figure 2. 176 Data analysis 177 Reads generated by DNBSEQ™ platform and Illumina MiSeq were merged using 178 VSEARCH 11 (v2.15.2_linux_x86_64) with the following parameters: --fastq_minovlen 10 --179 fastq_maxdiffs 15 --fastq_maxee 1 --fastq_allowmergestagger --fastq_qmaxout 92 --180 fastq_qmax 55. Since the sequencing depth of the DNBSEQ™ platform was approximately 181 225-fold higher (109,080,339 merged reads) than that of the Illumina MiSeq (485,342 182 merged reads), the merged DNBSEQ™ read pool was subsampled down to 485,342 reads. 183 The merged reads were discarded if the gap sequence (target region between probes) did not 184 match the target-specific probe sites, had an incorrect gap length, or had discrepancies in 185 three pool-specific nucleotides (2.7% of Illumina MiSeq merged reads and 4.5% of 186 preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for thisthis version posted March 12, 2026. ; https://doi.org/10.64898/2026.03.10.710732doi: bioRxiv preprint 10 DNBSEQ™-G99 merged reads). The error rate for each platform was determined by 187 comparing the gap sequence to the expected reference sequence. 188 All data were processed using proprietary pipelines. Python version 3.11.14 189 (https://www.python.org/downloads/release/python-31114/) and matplotlib library version 190 3.10.8 (https://matplotlib.org/3.10.8) were used to draw the figures. SciPy v1.17.0 191 (https://pypi.org/project/scipy/1.17.0/) library was used for Spearman correlations. 192

193 Nicking Loop™ libraries for circular and linear NGS platforms 194 To assess compatibility with circular NGS platforms, circular Nicking Loop™ libraries were 195 sequenced on the DNBSEQ™-G99. Nicking Loop™ was initially established using linear 196 libraries on Illumina MiSeq and therefore served as a benchmark for performance of the 197 circular libraries. 198 As the circular Nicking Loop™ method has been described in detail in a previous study10, 199 only the key differences relevant to platform compatibility are summarized here. In the first 200 step, the bridge oligonucleotide, probes, and the Loop are hybridized to linear target-specific 201 DNA. This hybridized intermediate is circularized, forming Nicking Loop™-converted 202 CssDNA that is further amplified by RCA to generate a long concatemer (Fig. 1A, steps 1-3). 203 To prepare a circular library, the concatemers are folded into a hairpin structure with a 204 nicking restriction site in the duplex stem, which is cleaved to produce monomers. The 205 monomers fold to create a circular molecule with a hairpin structure, forming a nick in the 206 stem. This nick is ligated, generating a circular library that can be used with different circular 207 sequencing platforms. For sequencing with the DNBSEQ™ platform, the libraries 208 additionally undergo DNA nanoball generation (Fig 1A, steps 4a–7a). 209 preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for thisthis version posted March 12, 2026. ; https://doi.org/10.64898/2026.03.10.710732doi: bioRxiv preprint 11 To prepare libraries for the Illumina platform, the RCA-generated concatemers were 210 subjected to limited-cycle PCR to introduce flowcell binding sequences and produce short 211 linear libraries compatible with bridge amplification (Fig 1A, steps 4b–7b). 212 213 Figure 1. Nicking Loop™ targeted library preparation method. (A) The conversion of linear DNA to circular DNA 214 and platform-specific processing steps for circular DNBSEQ™ and linear Illumina platform. The Nicking Loop™ construct 215 comprises a bridge oligonucleotide, the Loop oligonucleotide, and left and right target-specific probes that bind to target 216 DNA (1). Hybridized construct is circularized, forming Nicking Loop™-converted circular single-stranded DNA 217 (CssDNA; 2) that is amplified by rolling circle amplification (RCA). The RCA-amplified concatemers are used to prepare 218 1. 2. 3. 4.a 5.a 6.a 4.b 5.b 6.b Hybridization Nicking Loop™-converted circular single-stranded DNA Rolling circle amplification (RCA) Monomers Nicking Loop™ circular libraries DNBSEQ™ sequencing Limited-cycle PCR Nicking Loop™ linear libraries Illumina sequencing A B Sample-specifc index Restriction site Bridge-specific site preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for thisthis version posted March 12, 2026. ; https://doi.org/10.64898/2026.03.10.710732doi: bioRxiv preprint 12 both circular and linear libraries (3). For circular libraries, the concatemers are cleaved into monomers (4a). The monomers 219 fold and circularize, generating circular libraries (5a) that are sequenced with DNBSEQ™ platform (6a). For linear libraries, 220 RCA amplified concatemers are subjected to limited-cycle PCR (4b) to generate linear libraries (5b) that are sequenced with 221 Illumina platform (6b). (B) The Loop oligonucleotide consists of bridge-specific site required for incorporation into 222 hybridized construct, complementary restriction site for nicking endonuclease, and sample-specific index used for early 223 sample indexing. 224 Early sample indexing with Nicking Loop™ 225 The Loop oligonucleotide is an integral component of the Nicking Loop™ library preparation 226 and contains a sample-specific index (Fig. 1B). Upon incorporation of the Loop into the 227 hybridization construct with target DNA, the samples are tagged at an early stage of library 228 preparation. In this study, the early sample-specific index was used to demultiplex 229 sequencing reads generated by the DNBSEQ™-G99 (Supplementary Fig. 3), as the entire 230 structure of the Nicking Loop™ circular library was sequenced (Fig. 1A, step 5a; 231 Supplementary Fig. 1). In contrast, for linear libraries sequenced on the Illumina MiSeq, 232 sample-specific indices were introduced in the final step of library preparation workflow, as 233 the Loop structure was not sequenced by the Illumina MiSeq platform (Fig. 1A, steps 5b; 234 Supplementary Fig. 2). 235 UMI distribution across different Nicking Loop™ libraries and NGS platforms 236 The Nicking Loop™ design incorporates dual UMIs located within the left and right probe 237 arms flanking the target-specific region (Supplementary Fig 1). UMI distributions were 238 assessed across different Nicking Loop™ libraries and NGS platforms. To ensure direct 239 comparability, merged reads of the DNBSEQ™-G99 were subsampled to match the 240 sequencing depth of Illumina MiSeq (approximately 225-fold difference). Despite varying 241 sequencing depths, the general profiles of read and UMI coverage per probe were comparable 242 (Supplementary Fig. 3). 243 preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for thisthis version posted March 12, 2026. ; https://doi.org/10.64898/2026.03.10.710732doi: bioRxiv preprint 13 The total UMI count was 429,216 and 459,100 for the DNBSEQ™-G99 (subsampled) and 244 Illumina MiSeq, respectively. Singleton UMIs accounted for 97.36% and 97.03% of total 245 UMIs for the DNBSEQ™™-G99 and the Illumina MiSeq. The UMI distributions for each 246 probe were comparable across platforms (Fig. 2). In some cases, circular Nicking Loop™ 247 libraries sequenced on the DNBSEQ™-G99 platform yielded higher reads per UMI than 248 linear libraries sequenced on Illumina MiSeq, however, these differences were negligible 249 (0.099% of cases). 250 Beyond the matched-depth comparison of DNBSEQ™-G99, the UMI distribution was 251 assessed at native sequencing depth (109,080,339 merged reads). The distribution remained 252 dominated by UMIs with low read coverage (58.87% singleton UMIs), indicating high 253 molecular diversity (Supplementary Fig. 4). 254 255 Figure 2. Distribution of reads per unique molecular identifier (UMI) across Nicking Loop™ circular libraries sequenced on 256 the DNBSEQ™-G99 (blue) and linear libraries sequenced on the Illumina MiSeq (orange) for each probe in the synthetic 257 panel. 258 Mismatches introduced across different Nicking Loop™ libraries and NGS platforms 259 To determine the impact of platform-specific Nicking Loop™ steps and sequencing 260 chemistries on error rates, mismatches between the sequenced gap region and the expected 261 preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for thisthis version posted March 12, 2026. ; https://doi.org/10.64898/2026.03.10.710732doi: bioRxiv preprint 14

sequence were evaluated. Error rates were compared using depth-normalized 262 merged reads to ensure direct comparability between platforms. 263 Overall, fewer mismatches were observed for the circular Nicking Loop™ workflow 264 sequenced on the DNBSEQ™-G99 (99.74% correct reads) compared to linear Nicking 265 Loop™ workflow sequenced on the Illumina MiSeq platform (99.29% correct reads), 266 although the observed difference was small. Single base pair mismatches accounted for the 267 majority of all mismatches (0.25% and 0.70% of reads for DNBSEQ™-G99 and Illumina 268 MiSeq, respectively), while reads containing two or more mismatching base pairs were rare. 269 The error rate of DNBSEQ™-G99 at native sequencing depth was matching the depth-270 normalized results. At native sequencing depth, 99.76% of reads were correct, 0.23% of reads 271 contained a single base-pair mismatch, and the number of reads with ≥ 2 base-pair 272 mismatches was negligible (Supplementary table 1). 273 Table 1. Number of mismatches (bp) observed in gap sequence of merged reads across circular and linear Nicking Loop™ 274 workflows when compared to the expected reference gap sequence. Circular library was sequenced by DNBSEQ™-G99 and 275 linear library was sequenced with Illumina MiSeq. Reads obtained by DNBSEQ™-G99 were subsampled without replacement 276 to match the sequencing depth of Illumina MiSeq. 277 Platform Number of mismatches per gap sequence (bp) Number of merged reads Ratio DNBSEQ™-G99 (circular library) 0 463,396 99.74% 1 1,173 0.25% 2 7 0.0016% Illumina MiSeq (linear library) 0 472,074 99.29% 1 3,272 0.70% 2 24 0.0052% 278 preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for thisthis version posted March 12, 2026. ; https://doi.org/10.64898/2026.03.10.710732doi: bioRxiv preprint 15 VAF detection across different Nicking Loop™ libraries and NGS platforms 279 To assess consistency of VAF quantification across library preparation methods and 280 sequencing platforms, synthetic templates mimicking multiple VAFs were processed by 281 Nicking Loop™ and sequenced (Fig. 3). Sequencing results revealed a strong VAF 282 agreement for circular libraries sequenced on DNBSEQ™-G99 and linear libraries sequenced 283 on Illumina MiSeq (Spearman’s ρ = 0.939). The concordance indicates that platform-specific 284 downstream processing steps in the Nicking Loop™ workflow does not affect VAF 285 quantification. 286 287 Figure 3. Concordance of variant allele frequencies detected by Nicking Loop ™ library preparation workflow for the 288 DNBSEQ™ platform and the Illumina MiSeq. A strong agreement in VAF concordance was observed (Spearman’s ρ = 0.939) 289 despite platform-specific steps of Nicking Loop™ library preparation workflow. 290

291 Nicking Loop™ library preparation method provides flexibility for circular NGS platforms 292 by supporting sequencing of circular templates directly converted from linear target DNA or 293 following amplification. While direct circular sequencing offers a simplified protocol relative 294 to traditional library preparation, the amplified Nicking Loop™ libraries offer PCR-free 295 preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for thisthis version posted March 12, 2026. ; https://doi.org/10.64898/2026.03.10.710732doi: bioRxiv preprint 16 amplification with minimal bias. This study demonstrates that such amplified circular 296 libraries can be effectively sequenced on a circular NGS platforms and demultiplexed using 297 Loop index introduced in the initial step of the workflow, as illustrated by the results from the 298 DNBSEQ™ platform. Furthermore, the agreement observed between circular and linear 299 Nicking Loop™ libraries highlights the robustness and platform-independent read-out of the 300 method, expanding prior demonstrations on PacBio ONSO and Oxford Nanopore platforms10. 301 The robustness of the Nicking Loop™ method across library types and NGS platforms was 302 demonstrated by consistent UMI distributions, noise profiles, and reliable VAF detection. 303 UMI distributions across library types were highly comparable, with both libraries generating 304 more than 97% of singleton UMIs. Negligible deviations were observed, where the circular 305 library exhibited more reads per UMI than linear library, representing less than 0.1% of total 306 UMIs. This effect is likely caused by stochastic amplification of minor number of templates 307 in RCA, and it did not impact VAF detection. 308 The error profiles observed across the libraries were comparable, indicating that 309 platform-specific steps and sequencing chemistries have a negligible impact on the error 310 rates. Circular libraries sequenced on DNBSEQ™-G99 exhibited fewer errors, however, 311 these differences were within expected variability. Consistent with the observations, the 312 VAFs detected across circular and linear libraries were highly concordant 313 (Spearman’s ρ = 0.939), further supporting the robustness of the method. 314 Together, these results demonstrate that Nicking Loop™ provides a robust and reproducible 315 platform-independent read-out across multiple NGS platforms. It enables a native targeted 316 sequencing library preparation for circular DNBSEQ™. DNBSEQ™ workflows are 317 primarily optimized for whole-genome sequencing, and commonly used targeted protocols 318 often rely on conversion of PCR-enriched libraries originally designed for linear platforms 319 such as Illumina. In contrast, the Nicking Loop™ library is inherently compatible with 320 preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for thisthis version posted March 12, 2026. ; https://doi.org/10.64898/2026.03.10.710732doi: bioRxiv preprint 17 circular sequencing architecture, as it directly converts linear target DNA into circular 321 template. The method leverages RCA amplification to limit error propagation, workflow 322 complexity, and material loss associated with library conversion. Importantly, the Nicking 323 Loop™ supports targeted sequencing applications central to clinical use, including high-324 sensitivity somatic variant detection in oncology and focused gene panels for hereditary 325 disease-testing. 326 preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for thisthis version posted March 12, 2026. ; https://doi.org/10.64898/2026.03.10.710732doi: bioRxiv preprint 18 DATA AVAILABILITY 327 The data supporting the findings of this study are available upon reasonable request. 328 SUPPLEMENTARY DATA 329 Supplementary Data are available online. 330 AUTHOR CONTRIBUTIONS 331 S.A. – Conceptualization, Methodology, Formal analysis, Writing - Original Draft, 332 Visualization 333 A.K. – Conceptualization, Methodology, Software, Formal analysis, Data Curation, Writing - 334 Review & Editing, Visualization 335 T.H. – Conceptualization, Methodology, Writing - Review & Editing, Project administration 336 H.R. – Methodology, Investigation, Writing - Original Draft, Writing - Review & Editing, 337 N.L. – Methodology, Software, Formal analysis, Data Curation, Writing - Review & Editing 338 A.M. – Investigation, Writing - Review & Editing 339 T.R. – Investigation, Writing - Review & Editing 340 J.K. – Data Curation, Writing - Review & Editing 341 J.B. – Writing - Review & Editing 342 J.L. – Writing - Review & Editing 343 C.X. – Supervision, Writing - Review & Editing, 344 M.T. – Conceptualization, Methodology, Supervision, Project administration, Funding 345 acquisition, Writing - Review & Editing 346 J.P.P. – Conceptualization, Methodology, Investigation, Writing - Original Draft, 347 Supervision, Project administration 348 preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for thisthis version posted March 12, 2026. ; https://doi.org/10.64898/2026.03.10.710732doi: bioRxiv preprint 19

349 The authors would like to express their gratitude to Voima Ventures, (Helsinki, Finland), 350 Almaral (Kaarina, Finland), Avohoidon Tutkimussäätiö (Espoo, Finland) and Business 351 Finland (Helsinki, Finland) for support and funding. 352 DECLARATION OF GENERATIVE AI AND AI -ASSISTED TECHNOLOGIES IN 353 THE WRITING PROCESS 354 During the preparation of this work, the authors used ChatGPT 5.2 in order to refine the 355 language and improve readability in certain parts of the manuscript. After using this tool, the 356 authors reviewed and edited the content as needed and take full responsibility for the content 357 of the publication. 358 preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for thisthis version posted March 12, 2026. ; https://doi.org/10.64898/2026.03.10.710732doi: bioRxiv preprint 20

359 1. Wang M. Next-Generation Sequencing (NGS). Edited by Pan S and Tang J. Clinical 360 Molecular Diagnostics, Singapore, Springer, 2021, pp. 305–27 361 2. Ambardar S, Gupta R, Trakroo D, Lal R, Vakhlu J. High Throughput Sequencing: An 362 Overview of Sequencing Chemistry. Indian J Microbiol, 2016, 56:394–404 363 3. Next generation sequencing and beyond: a review of genomic sequencing methods | 364 Functional & Integrative Genomics | Springer Nature Link, n.d. 365 4. Lang J, Zhu R, Sun X, Zhu S, Li T, Shi X, Sun Y, Yang Z, Wang W, Bing P, He B, 366 Tian G. Evaluation of the MGISEQ-2000 Sequencing Platform for Illumina Target Capture 367 Sequencing Libraries. Front Genet, 2021, 12 368 5. Xu H, Wang Z, Sa S, Yang Y, Zhang X, Li D. Identification of novel compound 369 heterozygous variants of the ALMS1 gene in a child with Alström syndrome by whole 370 genome sequencing. Gene, 2024, 929:148827 371 6. Jeon SA, Park JL, Park S-J, Kim JH, Goh S-H, Han J-Y, Kim S-Y. Comparison 372 between MGI and Illumina sequencing platforms for whole genome sequencing. Genes 373 Genom, 2021, 43:713–24 374 7. Rao J, Luo H, An D, Liang X, Peng L, Chen F. Performance evaluation of structural 375 variation detection using DNBSEQ whole-genome sequencing. BMC Genomics, 2025, 376 26:299 377 8. Bewicke-Copley F, Arjun Kumar E, Palladino G, Korfi K, Wang J. Applications and 378 analysis of targeted genomic sequencing in cancer studies. Comput Struct Biotechnol J, 2019, 379 17:1348–59 380 9. Ghoreyshi N, Heidari R, Farhadi A, Chamanara M, Farahani N, Vahidi M, Behroozi 381 J. Next-generation sequencing in cancer diagnosis and treatment: clinical applications and 382 future directions. Discov Onc, 2025, 16:578 383 preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for thisthis version posted March 12, 2026. ; https://doi.org/10.64898/2026.03.10.710732doi: bioRxiv preprint 21 10. Nicking LoopTM: Streamlined circular DNA libraries for precision genomics, DNA 384 data storage and universal NGS read-out | bioRxiv, n.d. 385 https://www.biorxiv.org/content/10.1101/2025.05.08.652165v1.full. (accessed January 12, 386 2026) 387 11. Rognes T, Flouri T, Nichols B, Quince C, Mahé F. VSEARCH: a versatile open 388 source tool for metagenomics. PeerJ, 2016, 4:e2584 389 preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for thisthis version posted March 12, 2026. ; https://doi.org/10.64898/2026.03.10.710732doi: bioRxiv preprint