Publications

Papers, patents, and abstracts

For a full and up-to-date list, see my Google Scholar profile.

† indicates equal contribution.

Journal Articles

2025

Benchmarking of germline copy number variant callers from whole genome sequencing data for clinical applications.

De La Vega FM, Irvine SA, Anur P, Potts K, Kraft L, Torres R, Kang P, Truong S, Lee Y, Han S, Onuchic V, Han J.

Bioinformatics Advances, 5(1), vbaf071.

DOI
We benchmark several short-read WGS CNV detection tools using reference cell lines to inform their clinical use. Tools vary in sensitivity (7%–83%) and precision (1%–76%), with few meeting the sensitivity needed for clinical testing. Callers generally perform better for deletions (up to 88% sensitivity) than duplications (up to 47% sensitivity), with poor detection of duplications under 5 kb. For CNVs in genes commonly included in clinical panels, significantly improved sensitivity and precision were observed. DRAGEN v4.2 high-sensitivity CNV calls, post-processed with custom filters, achieved 100% sensitivity and 77% precision on the optimized gene panel after excluding recurring artifacts, which may support clinical use with orthogonal confirmation of reportable CNVs.
2023

Reproducible evaluation of transposable element detectors with McClintock 2 guides accurate inference of Ty insertion patterns in yeast.

Chen J, Basting PJ, Han S, Garfinkel DJ, Bergman CM.

Mobile DNA, 14, 8.

DOI
We report a completely re-implemented version of McClintock written in Python using Snakemake and Conda that improves installation, error handling, speed, stability, and extensibility. McClintock 2 now includes 12 short-read TE detectors, auxiliary pre-processing and analysis modules, interactive HTML reports, and a simulation framework to reproducibly evaluate the accuracy of component TE detectors. When applied to Saccharomyces cerevisiae, we find substantial variation in the ability of components to identify the precise locations of non-reference TE insertions, with RelocaTE2 showing the highest recall and precision in simulated data. Best-in-class predictors have sufficient resolution to reveal a dyad pattern of integration in nucleosome-bound regions upstream of yeast tRNA genes for Ty1, Ty2, and Ty4.
2022

Local assembly of long reads enables phylogenomics of transposable elements in a polyploid cell line.

Han S†, Dias GB†, Basting PJ, Viswanatha R, Perrimon N, Bergman CM.

Nucleic Acids Research, 50(21), e124.

DOI
Animal cell lines often undergo extreme genome restructuring, including polyploidy and segmental aneuploidy that can impede de novo whole-genome assembly. We developed a novel WGA-independent bioinformatics method called TELR that identifies, locally assembles, and estimates allele frequency of TEs from long-read sequence data. Application of TELR to a ~130x PacBio dataset for the tetraploid Drosophila S2R+ cell line revealed many haplotype-specific TE insertions that arose by transposition after initial cell line establishment and subsequent tetraploidization. Local assemblies from TELR also allowed phylogenetic analysis of paralogous TEs, revealing that proliferation of TE families in vitro can be driven by single or multiple source lineages.
2022

Ongoing transposition in cell culture reveals the phylogeny of diverse Drosophila S2 sublines.

Han S, Dias GB, Basting PJ, Nelson MG, Patel S, Marzo M, Bergman CM.

Genetics, 221(3), iyac077.

DOI
Cultured cells are widely used in molecular biology despite poor understanding of how cell line genomes change in vitro over time. We sequenced the genomes of 25 Drosophila S2 cell sublines and show that transposable element insertions provide abundant markers for the phylogenetic reconstruction of diverse sublines. Our analysis revealed dramatically different patterns of genome organization across sublines and supported a model of ongoing transposition dominated by episodic activity of a small number of retrotransposon families. We conclude that substantial genome evolution occurs during long-term Drosophila cell culture, which may impact the reproducibility of experiments that do not control for subline identity.
2022

A novel transposable element-based authentication protocol for Drosophila cell lines.

Mariyappa D, Rusch DB, Han S, Luhur A, Overton D, Miller DFB, Bergman CM, Zelhof AC.

G3: Genes, Genomes, Genetics, 12(2), jkab403.

DOI
Drosophila cell lines are used by researchers to investigate various cell biological phenomena, yet methods to authenticate them are lacking. We developed a PCR-based next-generation sequencing protocol to cluster cell lines based on the genome-wide distribution of a limited number of diagnostic TE families. The protocol reliably identified various Drosophila cell lines during double-blind testing, and showed minimal changes with respect to TE family distribution when cells are passaged for at least 50 times. We conclude that this approach can accurately identify and distinguish the numerous Drosophila cell lines available to the research community, thereby aiding reproducible cell culture research.
2021

Transposable element profiles reveal cell line identity and loss of heterozygosity in Drosophila cell culture.

Han S, Basting PJ, Dias GB, Luhur A, Zelhof AC, Bergman CM.

Genetics, 219(2), iyab113.

DOI
We leverage the classical observation that transposable elements (TEs) proliferate in cultured Drosophila cells to demonstrate that genome-wide TE insertion profiles can reveal the identity and provenance of Drosophila cell lines. We identify multiple cases where TE profiles clarify the origin of cell lines (Sg4, mbn2, and OSS_E) relative to published reports, and provide evidence that insertions from only a subset of LTR retrotransposon families are necessary to mark cell line identity. We also develop ngs_te_mapper2, a bioinformatics approach to detect TE insertions and estimate intra-sample allele frequencies in legacy WGS data, which revealed loss of heterozygosity as a mechanism shaping the unique TE profiles that identify Drosophila cell lines.
2017

Genomic analysis of P elements in natural populations of Drosophila melanogaster.

Bergman CM, Han S, Nelson MG, Bondarenko V, Kozeretska I.

PeerJ, 5, e3824.

DOI
We compared estimates of genomic P element content with gonadal dysgenesis phenotypes for isofemale strains from three worldwide populations of D. melanogaster. P element abundance is highly correlated across different bioinformatics approaches, but sensitive to method and filtering strategy. North American strains have fewer P elements but a higher proportion of full-length elements than European or African strains. However, geographic differences in P element abundance and structure are not associated with corresponding differences in dysgenesis. We conclude that genomic P element content alone is not sufficient to explain variation in gonadal dysgenesis across strains.

Application Notes & Research Articles

2026

Deliver haplotype-resolved, copy-aware small variant detection in paralogous regions within a single WGS workflow: enabled by Illumina TruPath Genome using Multi-Region Joint Detection.

Han S

Illumina Application Note, M-GL-04321.

Application note (PDF)
TruPath Genome encodes long-range molecular linkage directly on the flow cell during sequencing, and the DRAGEN Multi-Region Joint Detection (MRJD) algorithm uses that linkage to call haplotype-resolved, copy-aware small variants in paralogous regions from a standard short-read WGS workflow. Evaluated across 31 samples spanning eight paralogous loci with 98.2-99.9% sequence homology and sizes from 18 to 357 kb, the approach recovered the expected pathogenic finding in every sample, produced phased locus-wide haplotypes, and clarified allelic configuration, gene-pseudogene assignment, and CNV context. Results agreed with orthogonal methods including long-read sequencing, and were generated within three days of DNA extraction, replacing cascades of secondary assays such as MLPA and long-range PCR that typically add days to weeks. This is recovery of known findings in an evaluation set rather than a measure of diagnostic sensitivity, and the note records that CFC1, IKBKG and HYDIN were evaluated but not supported in DRAGEN v4.5. For research use only.
2024

Overcoming the challenges in PMS2 high-homology regions for improved detection of pathogenic variants associated with Lynch syndrome.

Illumina DRAGEN research team

Illumina Genomics Research article, June 17, 2024.

Article
PMS2 shares high sequence identity with its pseudogene PMS2CL, so reads from the two loci map ambiguously and pathogenic variants associated with Lynch syndrome are missed or assigned to the wrong copy. This article describes the multi-region joint detection (MRJD) approach to the problem, reporting recall of approximately 99.7% for SNVs and 97.1% for indels against a long-range PCR truth set, and detection of all expected pathogenic variants in 22 clinical samples. Recall here measures detection: in high-sensitivity mode an allele can be detected while its assignment to PMS2 or PMS2CL remains ambiguous, and calls from this short-read-only workflow are not phased.
2022

Genotyping of high homology HBA1 and HBA2 from Illumina whole-genome sequencing.

Han S, Onuchic V, Rossi M, Roller E, Cameron D.

Illumina Genomics Research article, September 8, 2022.

Article
The alpha-globin locus is one of the most structurally complex and clinically important regions of the genome, and its two near-identical genes defeat standard short-read genotyping. This article introduces the DRAGEN HBA caller, a targeted method that genotypes HBA1 and HBA2 from standard whole-genome sequencing. Calls showed high concordance with long-read sequencing and perfect Mendelian consistency across family trios, and applying the caller to a large, diverse population cohort recovered the elevated alpha-thalassemia carrier frequencies previously reported in African and South Asian populations.

Conference Abstracts

2026

O20: Benchmarking comprehensive variant detection in challenging genomic regions using Constellation mapped read technology.

Facio F, Johnson B, Onuchic VF, Han S, et al.

Genetics in Medicine Open, 4.

Abstract
2025

Accurate HBA and SMN genotyping within an integrated NGS workflow: a step toward accessible carrier screening.

Carroll A, Onuchic VF, Han S, et al.

Abstracts from the 58th European Society of Human Genetics (ESHG) Conference, European Journal of Human Genetics, 33, 947–948.

Abstract
2025

Constellation mapped read technology, a novel on-flow-cell library prep for both high-quality short reads and proximity information in one prep delivering an expanded genome.

Kaper F, Weir J, Bekritsky M, Han S, et al.

Abstracts from the 58th European Society of Human Genetics (ESHG) Conference, European Journal of Human Genetics, 33, 1160.

Abstract

Patents

2026

Targeted variant calling using target enrichment sequencing data.

Onuchic VF, Han S, Carroll A, Halpern A.

WO Patent WO2026073129A1.

Patent
2025

Two-copy allele detection.

Bernal CE, Han S, Onuchic VF.

WO Patent WO2025250794A1.

Patent
2024

Methods and systems for determining copy number variant genotypes.

Han S, Onuchic VF, Chen X.

WO Patent WO2024010812A2.

Patent
2024

Region-ambiguous variant detection.

Onuchic VF, Han S.

US Patent Application 18/674,538.

Patent
2024

Multi-region joint detection.

Han S, Andrews D, Onuchic VF.

US Patent Application 18/674,279.

Patent

Thesis

2021

Novel computational strategies for the analysis of transposable elements in Drosophila cell culture genomes.

Han S.

Ph.D. Dissertation, University of Georgia.

Dissertation (PDF) ProQuest