Biology subjects

Alkan, C.

Publications and source records attributed to Alkan, C..

At least 19 recordsLinked to original sources

bioRxiv2026

KENT: Fast Pathogen Screening with NVIDIA Jetson Cluster in Resource-Constrained Areas

Metagenomic sequencing enables broad pathogen detection without requiring prior assumptions about the organisms present, but the memory and computational demands of classifying large sequencing datasets against reference databases often lead analysis pipelines to rely on desktop or server-grade computing platforms. This limits practical use in disasterstricken and other resource-constrained environments, where portable sequencing may be available but reliable access to cloud or laboratory infrastructure is not. We present KENT, a portable edge-computing framework for metagenomic pathogen screening based on a cluster of NVIDIA Jetson Nano modules. KENT adapts CuCLARK-l for low-memory embedded execution and extends it with a custom software stack for distributed read classification, abundance estimation, split-run abundance merging, and report generation. Using an MPI-based controller/responder architecture, KENT supports local, offline analysis with sample-level parallelism across multiple nodes. We evaluate KENT using long read wastewater metagenomic samples with a custom pathogen database and the CAMI2 Marine long read dataset. In the wastewater setting, KENT achieves an average speedup of 3.1x over the CPU-based CLARK-l baseline while maintaining close agreement in classified read abundance estimates. On CAMI2 Marine, KENT achieves an average speedup of 2.68x and similarly preserves CLARK-l abundance profiles. Evaluation against the CAMI2 long read taxonomic ground truth yields a mean species-level precision of 58.55%, recall of 93.36%, F1 score of 71.74%, and L1 abundance distance of 0.731, closely matching CLARK-l. Among ground truth species represented in the reference database, KENT detects 99.86% on average. A controlled database-size experiment further shows little change in classification throughput across the tested database configurations, while large sequencing inputs remain constrained by the Jetson Nanos unified memory capacity. Overall, KENT demonstrates that GPU-accelerated metagenomic classification can be adapted to low-memory embedded hardware while preserving the biological profiling behavior of the CLARK-l baseline, supporting portable and offline pathogen screening in resource-constrained settings. AvailabilityKENT is available at https://github.com/BilkentCompGen/KENT

bioinformatics↗

bioRxiv2026

Repeated mutation of a GT92 glycosyltransferase gene confers antiviral resistance in two Caenorhabditis species

Host-pathogen interactions evolve rapidly within species, providing natural genetic resources for the identification of specific ecological interaction factors. We previously identified RNA viruses that infect the nematodes C. elegans and C. briggsae in a species-specific manner. Wild strains of both host species demonstrate ample variation in viral sensitivity. Specifically, the wild C. elegans strain MY10, despite carrying a deletion in a key immunity factor, was among the most resistant strains. Here we use recombinant inbred lines and pool-sequencing approaches to genetically map the major MY10 resistance locus, narrowing down its position by CRISPR/Cas9 mediated recombination and testing candidates by genome editing. A rare non-synonymous polymorphism in the gtnt-1 gene, encoding a putative glycosyltransferase of the GT92 family, causes resistance to viral infection in MY10. We find that viral resistance through gtnt-1 mutation occurred repeatedly in C. elegans, with diverse resistance alleles each remaining at low frequency (<1%). Furthermore, leveraging closely related C. briggsae strains differing in viral susceptibility, we demonstrate that repeated reduction-of-function alleles of the Cbr-gtnt-1 ortholog similarly impair viral infection and enhance host fitness upon infection. In conclusion, we found recurrent evolution in two host species of reduction-of-function alleles of the gtnt-1 orthologs, which repeatedly lead to viral resistance yet remain at low frequency. These repeated events provide a case of transient ecological adaptation to a pathogen through recurrent mutation of the same gene in two species. The low population frequencies of the resistant alleles point to a changing eco-evolutionary context that prevents their spread in populations, resulting in high allelic heterogeneity.

evolutionary biology↗

bioRxiv2026

FASTR: Reimagining FASTQ via Compact Image-inspired Representation

MotivationHigh-throughput sequencing (HTS) enables population-scale genomics but generates massive datasets, creating bottlenecks in storage, transfer, and analysis. FASTQ, the standard format for over two decades, stores one byte per base and one byte per quality score, leading to inefficient I/O, high storage costs, and redundancy. Existing compression tools can mitigate some issues, but often introduce costly decompression or complex dependency issues. ResultsWe introduce FASTR, a lossless, computation-native successor to FASTQ that encodes each nucleotide together with its base quality score into a single 8-bit value. FASTR reduces file size by at least 2x while remaining fully reversible and directly usable for downstream analyses. Applying general-purpose compression tools on FASTR consistently yields higher compression ratios, 2.47, 3.64, and 4.8x faster compression, and 2.34, 1.96, 1.75x faster decompression than on FASTQ across Illumina, HiFi, and ONT reads. FASTR is machine-learning-ready, allowing reads to be consumed directly as numerical vectors or image-like representations. We provide a highly parallel software ecosystem for FASTQ-FASTR conversion and show that FASTR integrates with existing tools, such as minimap2, with minimal interface changes and no performance overhead. By eliminating decompression costs and reducing data movement, FASTR lays the foundation for scalable genomics analyses and real-time sequencing workflows. Availability and Implementationhttps://github.com/ALSER-Lab/FASTR

bioinformatics↗

bioRxiv2026

GenCore: Genomic distance estimation using Locally Consistent Parsing

In the era of exponential data generation, a fast, consistent, and efficient string processing technique is necessary to represent extensive genomic data. One of the earliest string processing techniques, predating MinHash and minimizer-based sketching, is Locally Consistent Parsing (LCP). This technique partitions an input string and identifies short, exactly occurring substrings called cores, which collectively cover the input string while maintaining Partition and Labeling Consistency. The iterative application of LCP yields progressively longer cores in a compressed format, thereby substantially enhancing the efficiency of genomic sequence representation and subsequent downstream analysis. We have previously developed Lcptools as the first iterative implementation of LCP for the DNA alphabet and demonstrated its effectiveness in identifying cores with minimal collisions. Here, we introduce GO_SCPLOWENC_SCPLOWCO_SCPLOWOREC_SCPLOW, a computational method that leverages LCP cores for the first time to sketch and estimate genomic distances for closely related large genomes, and successfully reconstruct simulated progression trees. GO_SCPLOWENC_SCPLOWCO_SCPLOWOREC_SCPLOW also successfully recapitulates primate phylogeny using both telomere-totelomere (T2T) assemblies and the PacBio HiFi reads for assembly-free comparisons. AvailabilityGO_SCPLOWENC_SCPLOWCO_SCPLOWOREC_SCPLOW is available at https://github.com/BilkentCompGen/gencore

bioinformatics↗

bioRxiv2025

Robust software development practices improve citations of RNA-seq tools

RNA sequencing (RNA-seq) has emerged as an exemplary technology in biology and clinical applications, offering a crucial complement to other transcriptomic profiling protocols due to its high sensitivity, precision, and accuracy in characterizing transcriptomes. However, the rapid proliferation of RNA-seq tools necessitates the adoption of robust software development practices. Such development underscores the critical need to examine how RNA-seq tools are developed, maintained, and distributed; and whether the data they generate is reproducible as all of these factors are essential for ensuring software reliability, transparency, and trust in scientific findings. We conducted a comprehensive assessment of 434 RNA-seq tools developed between 2008 and 2024, categorizing them based on the type of analysis they perform. Our evaluation encompassed their software development and distribution methodologies, as well as the attributes contributing to their widespread adoption and dependability within the biomedical community, which were quantified by factors such as package manager availability, containerization, multithreading support, documentation quality, and inclusion of example datasets. Our findings establish the first documented positive association between rigorous software development practices and their adoption of published RNA-seq tools as measured by citations (Mann-Whitney U test, p-value = 4.9 x 10-26). By identifying key characteristics of widely adopted software, our findings guide developing robust and user-friendly RNA-seq tools, thereby reinforcing the call for rigorous community-wide standards.

bioinformatics↗

bioRxiv2025

Adaptive Increase of Amylase Gene Copy Number in Peruvians Driven by Potato-rich Diets

Summary/AbstractThe salivary amylase gene (AMY1) exhibits remarkable copy number variation linked to dietary shifts in human evolution. While global studies highlight its structural complexity and association with starch-rich diets, localized selection patterns remain under explored. Here, we analyzed AMY1 copy number in 3,723 individuals from 85 populations, revealing that Indigenous Peruvian Andean populations possess the highest AMY1 copy number globally. A genome-wide analysis showed significantly higher amylase copy numbers in Peruvian Andean genomes compared to closely related populations. Further, we identified positive selection (selection coefficient of 0.0124, log likelihood ratio of 11.1543) at the nucleotide level on a haplotype harboring at least five haploid AMY1 copies, with a Peruvian Andean-specific expansion coinciding with potato domestication ([~]6-10 kya). Using ultra-long-read sequencing, we demonstrated that previously described recombination-based mutational mechanisms drive the formation of high-copy AMY1 haplotypes observed in Andean population. Our study provides a framework for investigating structurally complex loci and their role in human dietary adaptation.

genetics↗

bioRxiv2024

Spherical Echo-Planar Time-resolved Imaging (sEPTI) for rapid 3D quantitative T2* and Susceptibility imaging

PurposeTo develop a 3D spherical EPTI (sEPTI) acquisition and a comprehensive reconstruction pipeline for rapid high-quality whole-brain submillimeter T2* and QSM quantification. MethodsFor the sEPTI acquisition, spherical k-space coverage is utilized with variable echo-spacing and maximum kx ramp-sampling to improve efficiency and incoherency when compared to existing EPTI approaches. For reconstruction, an iterative rank-shrinking B0 estimation and odd-even high-order phase correction algorithms were incorporated into the reconstruction to better mitigate artifacts from field imperfections. A physics-informed unrolled network was utilized to boost the SNR, where 1-mm and 0.75-mm isotropic whole-brain imaging were performed in 45 and 90 seconds, respectively. These protocols were validated through simulations, phantom, and in vivo experiments. Ten healthy subjects were recruited to provide sufficient data for the unrolled network. The entire pipeline was validated on additional 5 healthy subjects where different EPTI sampling approaches were compared. Two additional pediatric patients with epilepsy were recruited to demonstrate the generalizability of the unrolled reconstruction. ResultssEPTI achieved 1.4 x faster imaging with improved image quality and quantitative map precision compared to existing EPTI approaches. The B0 update and the phase correction provide improved reconstruction performance with lower artifacts. The unrolled network boosted the SNR, achieving high-quality T2* and QSM quantification with single average data. High-quality reconstruction was also obtained in the pediatric patient using this network. ConclusionsEPTI achieved whole-brain distortion-free multi-echo imaging and T2* and QSM quantification at 0.75 mm in 90 seconds which has the potential to be useful for wide clinical applications.

bioengineering↗

bioRxiv2024

Natural variation in infection specificityof Caenorhabditis briggsae isolates by two RNA viruses

Antagonistic relationships such as host-virus interactions potentially lead to rapid evolution and specificity in interactions. The Orsay virus is so far the only horizontal virus naturally infecting the nematode C. elegans. In contrast, several related RNA viruses infect its congener C. briggsae, including Santeuil (SANTV) and Le Blanc (LEBV) viruses. Here we focus on the hosts intraspecific variation in sensitivity to these two intestinal viruses. Many temperate-origin C. briggsae strains, including JU1264 and JU1498, are sensitive to both, while many tropical strains, such as AF16, are resistant to both. Interestingly, some C. briggsae strains exhibit a specific resistance, such as the HK104 strain, specifically resistant to LEBV. The viral sensitivity pattern matches the strains geographic and genomic relationships. The heavily infected strains mount a seemingly normal small RNA response that is insufficient to suppress viral infection, while the resistant strains show no small RNA response, suggesting an early block in viral entry or replication. We use a genetic approach from the host side to map genomic regions participating in viral resistance polymorphisms. Using Advanced Intercrossed Recombinant Inbred Lines (RILs) between virus-resistant AF16 and SANTV-sensitive HK104, we detect Quantitative Trait Loci (QTLs) on chromosomes IV and III. Building RILs between virus-sensitive JU1498 and LEBV-resistant HK104 followed by bulk segregant analysis, we identify a chromosome II QTL. In both cases, further introgressions of the regions confirmed the QTLs. This diversity provides an avenue for studying virus entry, replication, and exit mechanisms, as well as host-virus specificity and the host response to a specific virus infection.

genetics↗

bioRxiv2023

Pre-processing of paleogenomes: Mitigating reference bias and postmortem damage in ancient genome data

Ancient DNA analysis is subject to various technical challenges, including bias towards the reference allele ("reference bias"), postmortem damage (PMD) that confounds real variants, and limited coverage. Here, we conduct a systematic comparison of alternative approaches against reference bias and against PMD. To reduce reference bias, we either (a) mask variable sites before alignment or (b) align the data to a graph genome representing all variable sites. Compared to alignment to the linear reference genome, both masking and graph alignment effectively remove allelic bias when using simulated or real ancient human genome data, but only if sequencing data is available in FASTQ or unfiltered BAM format. Reference bias remains indelible in quality-filtered BAM files and in 1240K-capture data. We next study three approaches to overcome postmortem damage: (a) trimming, (b) rescaling base qualities, and (c) a new algorithm we present here, bamRefine, which masks only PMD-vulnerable polymorphic sites. We find that bamRefine is optimal in increasing the number of genotyped loci up to 20% compared to trimming and in improving accuracy compared to rescaling. We propose graph alignment coupled with bamRefine to minimise data loss and bias. We also urge the paleogenomics community to publish FASTQ files.

bioinformatics↗

bioRxiv2022

TargetCall: Eliminating the Wasted Computation in Basecalling via Pre-Basecalling Filtering

Basecalling is an essential step in nanopore sequencing analysis where the raw signals of nanopore sequencers are converted into nucleotide sequences, i.e., reads. State-of-the-art basecallers employ complex deep learning models to achieve high basecalling accuracy. This makes basecalling computationally-inefficient and memory-hungry; bottlenecking the entire genome analysis pipeline. However, for many applications, the majority of reads do no match the reference genome of interest (i.e., target reference) and thus are discarded in later steps in the genomics pipeline, wasting the basecalling computation. To overcome this issue, we propose TargetCall, the first fast and widely-applicable pre-basecalling filter to eliminate the wasted computation in basecalling. TargetCalls key idea is to discard reads that will not match the target reference (i.e., off-target reads) prior to basecalling. TargetCall consists of two main components: (1) LightCall, a lightweight neural network basecaller that produces noisy reads; and (2) Similarity Check, which labels each of these noisy reads as on-target or off-target by matching them to the target reference. TargetCall filters out all off-target reads before basecalling; and the highly-accurate but slow basecalling is performed only on the raw signals whose noisy reads are labeled as on-target. Our thorough experimental evaluations using both real and simulated data show that TargetCall 1) improves the end-to-end basecalling performance of the state-of-the-art basecaller by 3.31 x while maintaining high (98.88%) sensitivity in keeping on-target reads, 2) maintains high accuracy in downstream analysis, 3) precisely filters out up to 94.71% of off-target reads, and 4) achieves better performance, sensitivity, and generality compared to prior works. We freely open-source TargetCall to aid future research in pre-basecalling filtering at https://github.com/CMU-SAFARI/TargetCall.

bioinformatics↗

bioRxiv2022

BLEND: A Fast, Memory-Efficient, and Accurate Mechanism to Find Fuzzy Seed Matches in Genome Analysis

Generating the hash values of short subsequences, called seeds, enables quickly identifying similarities between genomic sequences by matching seeds with a single lookup of their hash values. However, these hash values can be used only for finding exact-matching seeds as the conventional hashing methods assign distinct hash values for different seeds, including highly similar seeds. Finding only exact-matching seeds causes either 1) increasing the use of the costly sequence alignment or 2) limited sensitivity. We introduce BLEND, the first efficient and accurate mechanism that can identify both exact-matching and highly similar seeds with a single lookup of their hash values, called fuzzy seed matches. BLEND 1) utilizes a technique called SimHash, that can generate the same hash value for similar sets, and 2) provides the proper mechanisms for using seeds as sets with the SimHash technique to find fuzzy seed matches efficiently. We show the benefits of BLEND when used in read overlapping and read mapping. For read overlapping, BLEND is faster by 2.4x-83.9x (on average 19.3x), has a lower memory footprint by 0.9x-14.1 x (on average 3.8x), and finds higher quality overlaps leading to accurate de novo assemblies than the state-of-the-art tool, minimap2. For read mapping, BLEND is fasterby 0.8x-4.1x (on average 1.7x) than minimap2. Source code is available at https://github.com/CMU-SAFARI/BLEND.

bioinformatics↗

bioRxiv2022

ECOLE: Learning to call copy number variants on whole exome sequencing data

Copy number variants (CNV) are shown to contribute to the etiology of several genetic disor{-}ders. Accurate detection of CNVs on whole exome sequencing (WES) data has been a long sought-after goal for use in clinics. This was not possible despite recent improvements in performance because algo{-}rithms mostly suffer from low precision and even lower recall on expert-curated gold standard call sets. Here, we present a deep learning-based somatic and germline CNV caller for WES data, named ECOLE. Based on a variant of the transformer architecture, the model learns to call CNVs per exon, using high-confidence calls made on matched WGS samples. We further train and fine-tune the model with a small set of expert calls via transfer learning. We show that ECOLE achieves high performance on human ex{-}pert labeled data for the first time with 68.7% precision and 49.6% recall. This corresponds to precision and recall improvements of 18.7% and 30.8% over the next best-performing methods, respectively. We also show that the same fine-tuning strategy using tumor samples enables ECOLE to detect RT-qPCR validated variations in bladder cancer samples without the need for a control sample. ECOLE is available at https://github.com/ciceklab/ECOLE.

bioinformatics↗

bioRxiv2021

CONGA: Copy number variation genotyping in ancient genomes and low-coverage sequencing data

AO_SCPLOWBSTRACTC_SCPLOWTo date, ancient genome analyses have been largely confined to the study of single nucleotide polymorphisms (SNPs). Copy number variants (CNVs) are a major contributor of disease and of evolutionary adaptation, but identifying CNVs in ancient shotgun-sequenced genomes is hampered by typical low coverage (<1 x) and short fragments (<80 bps), precluding standard CNV detection software to be effectively applied to ancient genomes. Here we present CONGA, tailored for genotyping CNVs at low coverage. Simulations and down-sampling experiments suggest that CONGA can genotype deletions >1 kbps with F-scores >0.75 at [≥]1x, and distinguish between heterozygous and homozygous states. We applied CONGA to genotype 10,002 outgroup-ascertained deletions across a heterogenous set of 71 ancient human genomes spanning the last 50,000 years, produced using variable experimental protocols. A fraction of these (21/71) display divergent deletion profiles unrelated to their population origin, but attributable to technical factors such as coverage and read length. The majority of the sample (50/71), despite originating from nine different laboratories and having coverages 0.44x-26x (median 4x) and read lengths 52-121 bp (median 69), exhibit coherent deletion frequencies. Across these 50 genomes, inter-individual genetic diversity measured using SNPs and CONGA-genotyped deletions are strongly correlated. CONGA-genotyped deletions also display purifying selection signatures, as expected. CONGA thus paves the way for systematic CNV analyses in ancient genomes, despite the technical challenges posed by low and variable genome coverage.

genomics↗

bioRxiv2021

AirLift: A Fast and Comprehensive Technique for Remapping Alignments between Reference Genomes

AirLift is the first read remapping tool that enables users to quickly and comprehensively map a read set, that had been previously mapped to one reference genome, to another similar reference. Users can then quickly run downstream analysis of read sets for each latest reference release. Compared to the state-of-the-art method for remapping reads (i.e., full mapping), AirLift reduces the overall execution time to remap read sets between two reference genome versions by up to 27.4x. We validate our remapping results with GATK and find that AirLift provides high accuracy in identifying ground truth SNP/INDEL variants. Code AvailabilityAirLift source code and readme describing how to reproduce our results are available at https://github.com/CMU-SAFARI/AirLift.

bioinformatics↗

bioRxiv2020

Polishing Copy Number Variant Calls on Exome Sequencing Data via Deep Learning

Accurate and efficient detection of copy number variants (CNVs) is of critical importance due to their significant association with complex genetic diseases. Although algorithms that use whole genome sequencing (WGS) data provide stable results with mostly-valid statistical assumptions, copy number detection on whole exome sequencing (WES) data shows comparatively lower accuracy. This is unfortunate as WES data is cost efficient, compact and is relatively ubiquitous. The bottleneck is primarily due to non-contiguous nature of the targeted capture: biases in targeted genomic hybridization, GC content, targeting probes, and sample batching during sequencing. Here, we present a novel deep learning model, DECoNT, which uses the matched WES and WGS data and learns to correct the copy number variations reported by any off-the-shelf WES-based germline CNV caller. We train DECoNT on the 1000 Genomes Project data, and we show that we can efficiently triple the duplication call precision and double the deletion call precision of the state-of-the-art algorithms. We also show that our model consistently improves the performance independent from (i) sequencing technology, (ii) exome capture kit and (iii) CNV caller. Using DECoNT as a universal exome CNV call polisher has the potential to improve the reliability of germline CNV detection on WES data sets.

bioinformatics↗

bioRxiv2019

A robust benchmark for germline structural variant detection

New technologies and analysis methods are enabling genomic structural variants (SVs) to be detected with ever-increasing accuracy, resolution, and comprehensiveness. Translating these methods to routine research and clinical practice requires robust benchmark sets. We developed the first benchmark set for identification of both false negative and false positive germline SVs, which complements recent efforts emphasizing increasingly comprehensive characterization of SVs. To create this benchmark for a broadly consented son in a Personal Genome Project trio with broadly available cells and DNA, the Genome in a Bottle (GIAB) Consortium integrated 19 sequence-resolved variant calling methods, both alignment- and de novo assembly-based, from short-, linked-, and long-read sequencing, as well as optical and electronic mapping. The final benchmark set contains 12745 isolated, sequence-resolved insertion and deletion calls [≥]50 base pairs (bp) discovered by at least 2 technologies or 5 callsets, genotyped as heterozygous or homozygous variants by long reads. The Tier 1 benchmark regions, for which any extra calls are putative false positives, cover 2.66 Gbp and 9641 SVs supported by at least one diploid assembly. Support for SVs was assessed using svviz with short-, linked-, and long-read sequence data. In general, there was strong support from multiple technologies for the benchmark SVs, with 90 % of the Tier 1 SVs having support in reads from more than one technology. The Mendelian genotype error rate was 0.3 %, and genotype concordance with manual curation was >98.7 %. We demonstrate the utility of the benchmark set by showing it reliably identifies both false negatives and false positives in high-quality SV callsets from short-, linked-, and long-read sequencing and optical mapping.

genomics↗

bioRxiv2018

Characterization of segmental duplications and large inversions using Linked-Reads

Many algorithms aimed at characterizing genomic structural variation (SV) have been developed since the inception of high-throughput sequencing. However, the full spectrum of SVs in the human genome is not yet assessed. Most of the existing methods focus on discovery and genotyping of deletions, insertions, and mobile elements. Detection of balanced SVs with no gain or loss of genomic segments (e.g., inversions) is particularly a challenging task. Long read sequencing has been leveraged to find short inversions but there is still a need to develop methods to detect large genomic inversions. Furthermore, currently there are no algorithms to predict the insertion locus of large interspersed segmental duplications.\n\nHere we propose novel algorithms to characterize large (>40Kbp) interspersed segmental duplications and (>80Kbp) inversions using Linked-Read sequencing data. Linked-Read sequencing provides long range information, where Illumina reads are tagged with barcodes that can be used to assign short reads to pools of larger (30-50 Kbp) molecules. Our methods rely on split molecule sequence signature that we have previously described [11]. Similar to the split read, split molecules refer to large segments of DNA that span an SV breakpoint. Therefore, when mapped to the reference genome, the mapping of these segments would be discontinuous. We redesign our earlier algorithm, VALOR, to specifically leverage Linked-Read sequencing data to discover large inversions and characterize interspersed segmental duplications. We implement our new algorithms in a new software package, called VALOR2.\n\nAvailabilityVALOR2 is available at https://github.com/BilkentCompGen/valor.

bioinformatics↗

bioRxiv2018

Discovery of tandem and interspersed segmental duplications using high throughput sequencing

MotivationSeveral algorithms have been developed that use high throughput sequencing technology to characterize structural variations. Most of the existing approaches focus on detecting relatively simple types of SVs such as insertions, deletions, and short inversions. In fact, complex SVs are of crucial importance and several have been associated with genomic disorders. To better understand the contribution of complex SVs to human disease, we need new algorithms to accurately discover and genotype such variants. Additionally, due to similar sequencing signatures, inverted duplications or gene conversion events that include inverted segmental duplications are often characterized as simple inversions; and duplications and gene conversions in direct orientation may be called as simple deletions. Therefore, there is still a need for accurate algorithms to fully characterize complex SVs and thus improve calling accuracy of more simple variants. ResultsWe developed novel algorithms to accurately characterize tandem, direct and inverted interspersed segmental duplications using short read whole genome sequencing data sets. We integrated these methods to our TARDIS tool, which is now capable of detecting various types of SVs using multiple sequence signatures such as read pair, read depth and split read. We evaluated the prediction performance of our algorithms through several experiments using both simulated and real data sets. In the simulation experiments, using a 30x coverage TARDIS achieved 96% sensitivity with only 4% false discovery rate. For experiments that involve real data, we used two haploid genomes (CHM1 and CHM13) and one human genome (NA12878) from the Illumina Platinum Genomes set. Comparison of our results with orthogonal PacBio call sets from the same genomes revealed higher accuracy for TARDIS than state of the art methods. Furthermore, we showed a surprisingly low false discovery rate of our approach for discovery of tandem, direct and inverted interspersed segmental duplications prediction on CHM1 (less than 5% for the top 50 predictions). AvailabilityTARDIS source code is available at https://github.com/BilkentCompGen/tardis, and a corresponding Docker image is available at https://hub.docker.com/r/alkanlab/tardis/ Contactfhormozd@ucdavis.edu and calkan@cs.bilkent.edu.tr

bioinformatics↗