Biology subjects

Layer, R. M.

Publications and source records attributed to Layer, R. M..

At least 19 recordsLinked to original sources

bioRxiv2026

Population-scale interpretation of RNA isoform diversity enabled by Isopedia

Alternative splicing generates extensive transcriptomic complexity, yet "novelty" is often inflated because of incomplete reference annotations, with 20-70% of transcripts in RNA-Seq studies labeled as novel. Isopedia provides an expandable data structure for reference-agnostic isoform annotation, which we demonstrate here through a population-scale catalog of 1,007 long-read datasets spanning 37 diverse biological contexts. By transitioning from reference-dependent to evidence-weighted annotation, Isopedia provides the frequency-based context necessary to distinguish stochastic noise from biologically active isoforms. In HG002 benchmarks, Isopedia reduced apparent isoform novelty by up to 26-fold, achieving a >95% annotation rate even for low-abundance isoforms typically missed by standard catalogs. The framework further supports systematic exploration of challenging loci such as pseudogenes and gene fusions. Isopedia transforms isoform discovery into a systematic interpretation of the human transcriptome, providing a critical foundation for clinical and functional RNA research. Isopedia is open source and freely available: https://github.com/zhengxinchang/isopedia.

bioinformatics↗

bioRxiv2024

STABIX: Summary statistic-based GWAS indexing and compression

Genome-Wide Association Studies (GWAS) are widely used to investigate the role of genetics in disease traits, but the resulting file sizes from these studies are large, posing barriers to efficient storage, sharing, and querying. This issue is especially important for biobanks like the UK Biobank that publish GWAS for thousands of traits, increasing the volume of data that must be effectively managed. Current compression and query methods reduce file sizes and allow for quick genomic position-based queries but do not provide utility for quickly finding loci based on their summary statistics. For example, finding all SNVs in a particular p-value range would require decompressing and scanning the whole file. We propose a new tool, STABIX, which introduces summary-statistic-based queries and improves upon the standard bgzip compression and tabix query tool in both compression ratio and decompression speed. When applied to ten GWAS files from PanUKBB, STABIX created smaller compressed data and indices than tabix for all files, where bgzip and tbi files were an average of 1.2 times the size of STABIX compressed files and indexes. In the same ten files, STABIX per gene decompression was, on average 7x faster than tabix per gene decompression, and achieved faster per gene decompression times for over 99% of nearly 20,000 genes.

bioinformatics↗

bioRxiv2024

GenoSiS: A Biobank-Scale Genotype Similarity Search Architecture for Creating Dynamic Patient-Match Cohorts

Many patients do not experience optimal benefits from medical advances because clinical research does not adequately represent them. While the diversity of biomedical research cohorts is improving, ensuring that individual patients are adequately represented remains challenging. We propose a new approach, GenoSiS, which leverages machine learning-based similarity search to dynamically find patient-matched cohorts across different populations quickly. These cohorts could serve as reference cohorts to improve a range of clinical analyses, including disease risk score calculations and dosage decisions. While GenoSiS focuses on finding genetic similarity within a biobank, our similarity search architecture can be extended to represent other medically relevant patient characteristics and search other biobanks.

bioinformatics↗

bioRxiv2024

PARMIK: PArtial Read Matching with Inexpensive K-mers

Environmental metagenomic sampling is instrumental in preparing for future pandemics by enabling early identification of potential pathogens and timely intervention strategies. Novel pathogens are a major concern, especially for zoonotic events. However, discovering novel pathogens often requires genome assembly, which remains a significant bottleneck. A robust metagenomic sampling that is directly searchable with new infection samples would give us a real-time understanding of outbreak origins dynamics. In this study, we propose PArtial Read Matching with Inexpensive K-mers (PARMIK), which is a search tool for efficiently identifying similar sequences from a patient sample (query) to a metagenomic sample (read). For example, at 90% identity between a query and a read, PARMIK surpassed BLAST, providing up to 21% higher recall. By filtering highly frequent k-mers, we reduced PARMIKs index size by over 50%. Moreover, PARMIK identified longer alignments faster than BLAST, peaking at 1.57x, when parallelizing across 32 cores.

genomics↗

bioRxiv2024

STIX: Long-reads based Accurate Structural Variation Annotation at Population Scale

The prioritization of Structural Variants (SV), which is needed to rank and identify potential pathogenic alleles, is still in its infancy. This is exemplified over gnomAD only being able to annotate 33.5% of GIAB SVs. To overcome this, we present the first long-read based annotation resource for both GRCh38 and CHM13-T2T reference using STIX. In contrast to previous methods, STIX indexes SV-informative long-reads themselves, can thus be easily extended and accurately annotate all SV types including insertions. STIX successfully annotated 95.9% of GIAB Tier1 SVs. STIX further improved cancer based SV prioritization by highlighting 3,563 SV from COSMIC being common in the population. We further showcase that mosaic SV can be independently gained and may be widely spread throughout the population. This highlights the need for accurate SV population frequency annotation to further facilitate the adoption of SV via long-read sequencing in medical research and clinical applications.

bioinformatics↗

bioRxiv2024

The effects of biological knowledge graph topology on embedding-based link prediction

Due to the limited information available about rare diseases and their causal variants, knowledge graphs are often used to augment our understanding and make inferences about new gene-disease connections. Knowledge graph embedding methods have been successfully applied to various biomedical link prediction tasks but have yet to be adopted for rare disease variant prioritization. Here, we explore the effect of knowledge graph topology on knowledge graph embedding link prediction performance and challenge the assumption that massively aggregating knowledge graphs is beneficial in deciphering rare disease cases and improving prediction outcomes. We find that using a filtered version of the Monarch knowledge graph with only 11% of the original size results in notably improved model predictive performance. Additionally, these findings suggest that successful KG optimization depends on selecting high-quality information rather than simply maximizing the amount of data included.

bioinformatics↗

bioRxiv2023

Hypothesis Generation For Rare and Undiagnosed Diseases Through Clustering and Classifying Time-Versioned Biological Ontologies

Rare diseases affect 1-in-10 people in the United States and despite increased genetic testing, up to half never receive a diagnosis. Even when using advanced genome sequencing platforms to discover variants, if there is no connection between the variants found in the patients genome and their phe-notypes in the literature, then the patient will remain undiagnosed. When a direct variant-phenotype connection is not known, putting a patients information in the larger context of phenotype relation-ships and protein-protein-interactions may provide an opportunity to find an indirect explanation. Databases such as STRING contain millions of protein-protein-interactions and HPO contains the relations of thousands of phenotypes. By integrating these networks and clustering the entities within we can potentially discover latent gene-to-phenotype connections. The historical records for STRING and HPO provide a unique opportunity to create a network time series for evaluating the cluster sig-nificance. Most excitingly, working with Childrens Hospital Colorado we provide promising hy-potheses about latent gene-to-phenotype connections for 38 patients with undiagnosed diseases. We also provide potential answers for 14 patients listed on MyGene2. Clusters our tool finds significant harbor 2.35 to 8.72 times as many gene-to-phenotypes edges inferred from known drug interactions than clusters find to be insignificant. Our tool, BOCC, is available as a web app and command line tool.

bioinformatics↗

bioRxiv2023

TEPEAK : A novel method for identifying and characterizing polymorphic transposable elements in non-model species populations

Transposable elements (TEs) replicate within genomes and are an active source of genetic variability in many species. Their role in immunity and domestication underscores their biological significance. However, analyzing TEs, especially within lesser-studied and wild populations, poses considerable challenges. To address this, we introduce TEPEAK, a simple and efficient approach to identify and characterize TEs in populations without any prior sequence or loci information. In addition to processing user-submitted genomes, TEPEAK integrates with the National Center for Biotechnology Information (NCBI) Sequence Read Archive (SRA) to increase cohort sizes or incorporate proximate species. Our application of TEPEAK to 256 horse genomes spanning 11 groups reaffirmed established genetic histories and highlighted disruptions in crucial genes. Some identified TEs were also detectable in species closely related to horses. TEPEAK paves the way for comprehensive genetic variation analysis in traditionally understudied populations by simplifying TE studies. TEPEAK is open-source and freely available at https://github.com/mrburke00/TEPEAK.

bioinformatics↗

bioRxiv2023

Embracing firefly flash pattern variability with data-driven species classification

Many nocturnally active fireflies use precisely timed bioluminescent patterns to identify mates, making them especially vulnerable to light pollution. As urbanization continues to brighten the night sky, firefly populations are under constant stress, and close to half of the species are now threatened. Ensuring the survival of firefly biodiversity depends on a large-scale conservation effort to monitor and protect thousands of populations. While species can be identified by their flash patterns, current methods require expert measurement and manual classification and are infeasible given the number and geographic distribution of fireflies. Here we present the application of a recurrent neural network (RNN) for accurate automated firefly flash pattern classification. Using recordings from commodity cameras, we can extract flash trajectories of individuals within a swarm and classify their species with a precision and recall of approximately seventy percent. In addition to scaling population monitoring, automated classification provides the means to study firefly behavior at the population level. We employ the classifier to measure and characterize the variability within and between swarms, unlocking a new dimension of their behavior. Our method is open source, and deployment in community science applications could revolutionize our ability to monitor and understand firefly populations.

ecology↗

bioRxiv2021

Mining Thousands of Genomes to Classify Somatic and Pathogenic Structural Variants

Structural variants (SVs) are associated with cancer progression and Mendelian disorders, but challenges with estimating SV frequency remain a barrier to somatic and de novo classification. In particular, variability in filtering and variant calling heuristics limit our ability to use SV catalogs from large cohorts. We present a method to index and search the raw alignments from thousands of samples that overcomes these limitations and supports robust SV analysis.

bioinformatics↗

bioRxiv2020

Samplot: A Platform for Structural Variant Visual Validation and Automated Filtering

Visual validation is an essential step to minimize false positive predictions resulting from structural variant (SV) detection. We present Samplot, a tool for quickly creating images that display the read depth and sequence alignments necessary to adjudicate purported SVs across multiple samples and sequencing technologies, including short, long, and phased reads. These simple images can be rapidly reviewed to curate large SV call sets. Samplot is easily applicable to many biological problems such as prioritization of potentially causal variants in disease studies, family-based analysis of inherited variation, or de novo SV review. Samplot also includes a trained machine learning package that dramatically decreases the number of false positives without human review. Samplot is available via the conda package manager or at https://github.com/ryanlayer/samplot. ContactRyan Layer, Ph.D., Assistant Professor, University of Colorado Boulder, ryan.layer@colorado.edu.

genetics↗

bioRxiv2020

Learning What a Good Structural Variant Looks Like.

Structural variations (SVs) are an important class of genetic mutations, yet SV detectors still suffer from high false-positive rates. In many cases, humans can quickly determine whether a putative SV is real by merely looking at a visualization of the SVs coverage profile. To that end, we developed Samplot-ML, a convolutional neural network (CNN) trained to genotype genomic deletions using Samplot visualizations that incorporate various forms of evidence such as genome coverage, discordant pairs, and split reads. Using Samplot-ML, we were able to reduce false positives by 47% while keeping 97% of true positives on average across several test samples.

bioinformatics↗

bioRxiv2020

The structural variation landscape in 492 Atlantic salmon genomes

Structural variants (SVs) are a major source of genetic and phenotypic variation, but remain challenging to accurately type and are hence poorly characterized in most species. We present an approach for reliable SV discovery in non-model species using whole genome sequencing and report 15,483 high-confidence SVs in 492 Atlantic salmon (Salmo salar L.) sampled from a broad phylogeographic distribution. These SVs recover population genetic structure with high resolution, include an active DNA transposon, widely affect functional features, and overlap more duplicated genes retained from an ancestral salmonid autotetraploidization event than expected. Changes in SV allele frequency between wild and farmed fish indicate polygenic selection on behavioural traits during domestication, targeting brain-expressed synaptic networks linked to neurological disorders in humans. This study offers novel insights into the role of SVs in genome evolution and the genetic architecture of domestication traits, along with resources supporting reliable SV discovery in non-model species.

genomics↗

bioRxiv2020

Regulatory Sharing Between Estrogen Receptor α Bound Enhancers

The human genome encodes an order of magnitude more gene expression enhancers than promoters, suggesting that most genes are regulated by the combined action of multiple enhancers. We have previously shown that neighboring estrogen-responsive enhancers, which are approximately 5,000 basepairs apart, exhibit complex synergistic contributions to the production of an estrogenic transcriptional response. Here we sought to determine the molecular underpinnings of the observed enhancer cooperativity. We generated genetic deletions of individual estrogen receptor (ER) bound enhancers and found that enhancers containing full estrogen response element (ERE) motifs control ER binding at neighboring sites, while enhancers with pre-existing histone acetylation/accessibility confer a permissible chromatin environment to the neighboring enhancers. Genome engineering revealed that a cluster of two enhancers with half EREs could not compensate for the lack of a full ERE site within the cluster. In contrast, two enhancers with full EREs produced a transcriptional response greater than the wild-type locus. By swapping genomic sequences between enhancers, we found that the genomic location in which a full ERE resides strongly influences enhancer activity. Our results lead to a model in which a full ERE is required for ER recruitment, but the presence of a pre-existing active chromatin environment within an enhancer cluster is also needed in order for estrogen-driven gene regulation to occur.

genetics↗

bioRxiv2018

Mapping and characterization of structural variation in 17,795 deeply sequenced human genomes

A key goal of whole genome sequencing (WGS) for human genetics studies is to interrogate all forms of variation, including single nucleotide variants (SNV), small insertion/deletion (indel) variants and structural variants (SV). However, tools and resources for the study of SV have lagged behind those for smaller variants. Here, we used a cloud-based pipeline to map and characterize SV in 17,795 deeply sequenced human genomes from common disease trait mapping studies. We publicly release site-frequency information to create the largest WGS-based SV resource to date. On average, individuals carry 2.9 rare SVs that alter coding regions, which affect the dosage or structure of 4.2 genes and account for 4.0-11.2% of rare high-impact coding alleles. Based on a computational model, we estimate that SVs account for 17.2% of rare alleles genome-wide whose predicted deleterious effects are equivalent to loss-of-function (LoF) coding alleles; ~90% of such SVs are non-coding deletions (mean 19.1 per genome). We report 158,991 ultra-rare SVs and show that ~2% of individuals carry ultra-rare megabase-scale SVs, nearly half of which are balanced and/or complex rearrangements. Finally, we exploit this resource to infer the dosage sensitivity of genes and non-coding elements, revealing strong trends related to regulatory element class, conservation and cell-type specificity. This work will help guide SV analysis and interpretation in the era of WGS.

genomics↗

bioRxiv2018

svtools: population-scale analysis of structural variation

SummaryLarge-scale human genetics studies are now employing whole genome sequencing with the goal of conducting comprehensive trait mapping analyses of all forms of genome variation. However, methods for structural variation (SV) analysis have lagged far behind those for smaller scale variants, and there is an urgent need to develop more efficient tools that scale to the size of human populations. Here, we present a fast and highly scalable software toolkit (svtools) and cloud-based pipeline for assembling high quality SV maps - including deletions, duplications, mobile element insertions, inversions, and other rearrangements - in many thousands of human genomes. We show that this pipeline achieves similar variant detection performance to established per-sample methods (e.g., via LUMPY), while providing fast and affordable joint analysis at the scale of [≥]100,000 genomes. These tools will help enable the next generation of human genetics studies.\n\nAvailability and Implementationsvtools is implemented in Python and freely available (MIT) from https://github.com/hall-lab/svtools.\n\nContactihall@wustl.edu

bioinformatics↗

bioRxiv2018

SV-plaudit: A cloud-based framework for manually curating thousands of structural variants

SV-plaudit is a framework for rapidly curating structural variant (SVs) predictions. For each SV, we generate an image that visualizes the coverage and alignment signals from a set of samples. Images are uploaded to our cloud framework where users assess the quality of each image using a client-side web application. Reports can then be generated as a tab-delimited file or annotated VCF. As a proof of principle, nine researchers collaborated for one hour to evaluate 1,350 SVs each. We anticipate that SV-plaudit will become a standard step in variant calling pipelines and the crowd-sourced curation of other biological results.\n\nCode available at https://github.com/jbelyeu/SV-plaudit\n\nDemonstration video available at https://www.youtube.com/watch?v=ono8kHMKxDs

bioinformatics↗

bioRxiv2017

A map of constrained coding regions in the human genome.

Deep catalogs of genetic variation collected from many thousands of humans enable the detection of intraspecies constraint by revealing coding regions with a scarcity of variation. While existing techniques summarize constraint for entire genes, single metrics cannot capture the fine-scale variability in constraint within each protein-coding gene. To provide greater resolution, we have created a detailed map of constrained coding regions (CCRs) in the human genome by leveraging coding variation observed among 123,136 humans from the Genome Aggregation Database (gnomAD). The most constrained coding regions in our map are enriched for both pathogenic variants in ClinVar and de novo mutations underlying developmental disorders. CCRs also reveal protein domain families under high constraint, suggest unannotated or incomplete protein domains, and facilitate the prioritization of previously unseen variation in studies of disease. Finally, a subset of CCRs with the highest constraint likely exist within genes that cause yet unobserved human phenotypes owing to strong purifying selection.

genomics↗