Biology subjects

Sundar, H.

Publications and source records attributed to Sundar, H..

4 recordsLinked to original sources

bioRxiv2026

Generation of three induced pluripotent stem cell lines from an immune checkpoint inhibitor-induced myocarditis patient and controls

Immune checkpoint inhibitor-associated myocarditis (ICIM) is an uncommon but potentially fatal inflammatory heart disease triggered by cancer immunotherapy, with up to 40% mortality. The underlying mechanisms are still elusive, partly due to the lack of appropriate human disease models. Here, we report the generation of three induced pluripotent stem cell (iPSC) lines derived from an ICIM patient, an ICI-treated patient without myocarditis, and a healthy donor. These lines exhibit typical pluripotent stem cell morphology, express pluripotency markers, maintain normal karyotypes, and differentiate into derivatives of the three germ layers, providing a valuable platform for mechanistic studies and therapeutic discovery.

Developmental Biology↗

bioRxiv2026

LAMBDA: A Prophage Detection Benchmark for Genomic Language Models

Transformer-based genomic sequence models represent an emerging frontier in computational biology. Yet, their embeddings have not yet shown the same level of predictive power as natural and protein language models, highlighting a gap between current implementations and theoretical promise. Existing benchmarks for DNA language models primarily focus on classifying regulatory elements in eukaryotic genomes, leaving open the fundamental question of whether these models learn sequence-level features across whole genomes. We introduce LAMBDA, a benchmark designed to rigorously evaluate genome language model embeddings through phage-bacteria sequence discrimination across four categories of increasing complexity: probing tasks, fine-tuning assessments, diagnostic tests, and genome-wide prophage detection. Our comprehensive analysis of current genomic language models provides insight into the importance of training data selection relative to model size, the need for domain-specific training, and the capabilities and limitations of genomic language models for detecting prophage sequences. This benchmark represents a challenging genomic annotation task in the bacterial domain and addresses a key computational problem with direct relevance to microbiology and medicine.

genomics↗

bioRxiv2024

A Comparison of Tokenization Impact in AttentionBased and State Space Genomic Language Models

Genomic language models have recently emerged as a new method to decode, interpret, and generate genetic sequences. Existing genomic language models have utilized various tokenization methods, including character tokenization, overlapping and non-overlapping k-mer tokenization, and byte-pair encoding, a method widely used in natural language models. Genomic sequences differ from natural language because of their low character variability, complex and overlapping features, and inconsistent directionality. These features make sub-word tokenization in genomic language models significantly different from both traditional language models and protein language models. This study explores the impact of tokenization in genomic language models by evaluating their downstream performance on forty-four classification fine-tuning tasks. We also perform a direct comparison of byte pair encoding and character tokenization in Mamba, a state-space model. Our results indicate that character tokenization outperforms sub-word tokenization methods on tasks that rely on nucleotide level resolution, such as splice site prediction and promoter detection. While byte-pair tokenization had stronger performance on the SARS-CoV-2 variant classification task, we observed limited statistically significant differences between tokenization methods on the remaining downstream tasks.

bioinformatics↗

bioRxiv2020

Virulent disease epidemics can increase host density by depressing foraging of hosts

All else equal, parasites that harm host fitness should depress densities of their hosts. However, parasites that alter host traits may increase host density via indirect ecological interactions. Here, we show how depression of infected host foraging rate can produce such a hydra effect. Using a foraging assay, we quantified reduced foraging rates of a zooplankton host infected with a virulent fungal parasite. We then parameterized a dynamical model of hosts, parasites, and resources with this foraging function, showing how foraging depression can create a hydra effect. Mathematically, the hydra arose when increased resource productivity exceeded any increase in resource consumption per host. Therefore, the foraging-mediated hydra effect more likely emerged (1) for hosts which strongly control logistic-like resources and (2) during larger epidemics of moderately virulent parasites. We then analyzed epidemics from 13 fungal epidemics in nature. We found evidence for a foraging-mediated hydra effect: large outbreaks depressed foraging rate and correlated with increased densities of both algae and hosts. Therefore, depression of foraging rate of infected hosts can produce higher host densities even during epidemics of parasites that increase host mortality. Such hydras might prevent collapse of host populations but also could produce higher densities of infected hosts.Competing Interest StatementThe authors have declared no competing interest.View Full Text

ecology↗