Biology subjects

Steely, C. J.

Publications and source records attributed to Steely, C. J..

5 recordsLinked to original sources

bioRxiv2025

Systematic contextual biases in SegmentNT relevant to all nucleotide transformer models

Recent advances in large language models (LLMs) have extended to genomic applications, yet model robustness relative to context is unclear. Here, we demonstrate two intrinsic biases (input sequence length and nucleotide position) affecting SegmentNT results, a model included with the Nucleotide Transformer that provides nucleotide-level predictions of biological features. We demonstrate that nucleotide position within the input sequence (beginning, middle, or end) alters the nature of SegmentNTs raw prediction probabilities, which can be standardized to improve prediction consistency. While longer input sequence length improves model performance, diminishing returns suggest a surprisingly small input length of [~]3,072 nucleotides might be sufficient for many applications. We further identify a 24-nucleotide periodic oscillation in SegmentNTs prediction probabilities, revealing an intrinsic bias potentially linked to the models training tokenization (6-mers) and architecture. We identify potential approaches to account for these biases and provide generalizable insights for utilizing nucleotide-resolution functional prediction models.

bioinformatics↗

bioRxiv2024

A familial, telomere-to-telomere reference for human de novo mutation and recombination from a four-generation pedigree

Using five complementary short- and long-read sequencing technologies, we phased and assembled >95% of each diploid human genome in a four-generation, 28-member family (CEPH 1463) allowing us to systematically assess de novo mutations (DNMs) and recombination. From this family, we estimate an average of 192 DNMs per generation, including 75.5 de novo single-nucleotide variants (SNVs), 7.4 non-tandem repeat indels, 79.6 de novo indels or structural variants (SVs) originating from tandem repeats, 7.7 centromeric de novo SVs and SNVs, and 12.4 de novo Y chromosome events per generation. STRs and VNTRs are the most mutable with 32 loci exhibiting recurrent mutation through the generations. We accurately assemble 288 centromeres and six Y chromosomes across the generations, documenting de novo SVs, and demonstrate that the DNM rate varies by an order of magnitude depending on repeat content, length, and sequence identity. We show a strong paternal bias (75-81%) for all forms of germline DNM, yet we estimate that 17% of de novo SNVs are postzygotic in origin with no paternal bias. We place all this variation in the context of a high-resolution recombination map ([~]3.5 kbp breakpoint resolution). We observe a strong maternal recombination bias (1.36 maternal:paternal ratio) with a consistent reduction in the number of crossovers with increasing paternal (r=0.85) and maternal (r=0.65) age. However, we observe no correlation between meiotic crossover locations and de novo SVs, arguing against non-allelic homologous recombination as a predominant mechanism. The use of multiple orthogonal technologies, near-telomere-to-telomere phased genome assemblies, and a multi-generation family to assess transmission has created the most comprehensive, publicly available "truth set" of all classes of genomic variants. The resource can be used to test and benchmark new algorithms and technologies to understand the most fundamental processes underlying human genetic variation.

genomics↗

bioRxiv2021

The Mutational Dynamics of Short Tandem Repeats in Large, Multigenerational Families

Short tandem repeats (STRs) are tandemly repeated sequences of 1-6 bp motifs. STRs compose approximately 3% of the genome, and mutations at STR loci have been linked to dozens of human diseases including amyotrophic lateral sclerosis, Friedreich ataxia, Huntington disease, and fragile X syndrome. Improving our understanding of these mutations would increase our knowledge of the mutational dynamics of the genome and may uncover additional loci that contribute to disease. Here, to estimate the genome-wide pattern of mutations at STR loci, we analyzed blood-derived whole-genome sequencing data for 544 individuals from 29 three-generation CEPH pedigrees. These pedigrees contain both sets of grandparents, the parents, and an average of 9 grandchildren per family. Using HipSTR we identified de novo STR mutations in the 2nd generation of these pedigrees. Analyzing ~1.6 million STR loci, we estimate the empircal de novo STR mutation rate to be 5.24*10-5 mutations per locus per generation. We find that perfect repeats mutate ~2x more often than imperfect repeats. De novo STRs are significantly enriched in Alu elements (p < 2.2e-16). Approximately 30% of STR mutations occur within Alu elements, which compose only ~11% of the genome, and ~10% are found in LINE-1 insertions, which compose ~17% of the genome. Phasing these de novo mutations to the parent of origin shows that parental transmission biases vary among families. We estimate the average number of de novo genome-wide STR mutations per individual to be ~85, which is similar to the average number of observed de novo single nucleotide variants.

genomics↗

bioRxiv2021

MFN2 Influences Amyotrophic Lateral Sclerosis Pathology

ObjectiveTo better understand the pathology of amyotrophic lateral sclerosis, we used sequence data from patients seen at the University of Utah to identify novel disease-associated loci. We utilized both in vitro and in vivo studies to determine the biological effect of patient mutations in MFN2. MethodSequence data for a total of 140 patients were run through VAAST and Phevor to determine genes that were more burdened with rare, nonsynonymous variants compared to control longevity cohort. Variants identified in MFN2 were expressed in Mfn2 knockout cells to determine if mutant MFN2 could rescue mitochondrial morphology defects. We identified additional rare, nonsynonymous variants in MFN2 in ALSdb that were expressed in knockout mouse embryonic fibroblasts (MEFs). Membrane potential was measured to quantify mitochondrial health upon mutant MFN2 expression. mfn2 knockout zebrafish were used to examine movement compared to wildtype and protein aggregation in brain. ResultsMFN2 mutations identified in ALS patients from our University of Utah cohort and ALSdb were defective in rescuing morphological defects in Mfn2 knockout MEFs. Selected mutants showed decreased membrane potential compared to wildtype MFN2 expression. Zebrafish heterozygous and homozygous for loss of mfn2 showed increased TDP-43 levels in their hindbrain and cerebellum. ConclusionIn total, 21 rare, deleterious mutations in MFN2 were tested in Mfn2 knockout MEFs. Mutant MFN2 expression was not able to rescue the knockout phenotype, though at differing degrees of severity. Decreased membrane potential also argues for inhibited mitochondrial function. Increased TDP-43 levels in mutant zebrafish illustrates MFN2s function in ALS pathology. MFN2 variants influence ALS pathology and highlight the importance of mitochondria in neurodegeneration.

genetics↗

bioRxiv2020

Mobile Element Insertions and Associated Structural Variants in Longitudinal Breast Cancer Samples

While mobile elements are largely inactive in healthy somatic tissues, increased activity has been found in cancer tissues, with significant variation among different cancer types. In addition to insertion events, mobile elements have also been found to mediate many structural variation events in the genome. Here, to better understand the timing and impact of mobile element insertions and structural variants involving existing mobile elements in cancer, we examined their activity in longitudinal samples of four metastatic breast cancer patients. With whole-genome sequencing data from multiple timepoints through tumor progression, we used mobile element detection software followed by visual confirmation of the insertions. We identified 11 mobile element insertions or structural variants involving existing elements and found that the majority of these occurred early in tumor progression. Two of the identified insertions were SVA elements, which have rarely been found in previous cancer studies. Most of the variants appear to impact intergenic regions; however, we identified a translocation interrupting MAP2K4 involving Alu elements and a deletion in YTHDF2 involving mobile elements that likely inactivate reported tumor suppressor genes. The high variant allele fraction of the MAP2K4 translocation, the loss of the other copy of MAP2K4, the recurrent loss-of-function mutations found in this gene in other human cancers, and the important function of MAP2K4 indicate that this translocation is potentially a driver mutation. Overall, using a unique longitudinal dataset, we find that most variants are likely passenger mutations in the four patients we examined, but some variants impact tumor progression.

genomics↗