Biology subjects

Viswanathan, K.

Publications and source records attributed to Viswanathan, K..

9 recordsLinked to original sources

bioRxiv2025

BioToken and BioFM - Biologically-Informed Tokenization Enables Accurate and Efficient Genomic Foundation Models

Genomic variation underlies human phenotypic diversity, disease susceptibility, and evolutionary adaptation. Although large-scale genomic sequencing has transformed our ability to map genetic variation, accurately modeling and interpreting this data remains a central challenge; while genomic foundation models (GFMs) are a promising approach, they suffer from fundamental limitations. Current GFMs typically treat DNA simplistically as nucleotide-only sequences, overlooking critical biological context, such as genomic annotations, regulatory elements, and functional contexts central to genomic interpretation. Here, we introduce BioToken, a modular and extendable tokenization framework designed to encode genomic variants and biologically relevant region annotations directly into genomic representations. By utilizing intrinsic inductive biases, BioToken facilitates meaningful representation learning and generalization across diverse molecular phenotypes, such as gene expression, alternative splicing, and variant pathogenicity prediction. Built on BioToken, our genomic foundation model, BioFM, achieves competitive or superior results relative to specialized models (e.g., Enformer, SpliceTransformer) and GFMs up to 7B parameters across a comprehensive suite of genomic benchmarks, including noncoding pathogenicity, expression modulation, sQTL prediction, and long-range genomic interactions. Notably, BioFM achieves state-of-the-art performance with significantly fewer parameters (265M), substantially reducing training costs and computational requirements. Our findings highlight the substantial advantages of integrating biologically-informed inductive biases into genomic foundation modeling, providing a robust and accessible path forward in genomics. We provide our code and model checkpoints to support further research in this direction.

genomics↗

bioRxiv2025

Biopolymers can be used for consolidation of lunar and Martian regoliths

Creating long-term extraterrestrial habitats to support continuous human occupation requires structures that can bear load, while remaining resilient to large temperature fluctuations. Exploiting locally available regolith is part of a paradigm termed in situ resource utilization (ISRU) that advocates minimal dependence on terrestrial supplies via local resource exploitation for construction. In this work, we report on a consolidation method for making load-bearing structures out of lunar and Martian regolith simulants using biopolymer binders. We quantify the compressive and tensile strength of the final brick-like structures, termed biopolymer regolith composites, and investigate their dependence on biopolymer concentration and chemistry. A pressure-assisted die casting process is developed that can produce bricks with compressive strengths of [~] 15 MPa and tensile strengths of around 2 MPa, from regolith simulants. We finally report on the performance of these bricks under large temperature and pressure variations and find that their strengths remain unchanged. We believe that this makes biopolymer regolith composite bricks suitable candidates for use in harsh extraterrestrial environments, while adhering to the guidelines of the ISRU paradigm. Keyword: Lunar regolith; Martian regolith; Space habitation; Extraterrestrial construction; Biopolymers; Regolith simulant

bioengineering↗

bioRxiv2024

Genomic Foundationless Models: Pretraining Does Not Promise Performance

AO_SCPLOWBSTRACTC_SCPLOWThe success of Large Language Models has inspired the development of Genomic Foundation Models (GFMs) through similar pretraining techniques. However, the relationship between pretraining performance and effectiveness in downstream genomic tasks remains unclear. Additionally, the high computational cost of pretraining raises questions about its cost-efficiency. To assess the usefulness of pretraining in genomics, we evaluated seven different GFMs across 52 diverse genomic tasks, comparing them to their counterparts with randomly initialized weights. Across benchmarks, we find that randomly initialized models provide surprisingly strong baselines and tokenizer and architecture choices strongly shape both these baselines and the gains from pretraining. Specifically, character-token models often match or exceed the performance of larger pretrained k-mer or BPE models, whereas subword models appear to benefit from pretraining. We also find that the evaluated GFMs fail to capture clinically relevant genetic mutations, with embeddings and log-likelihood ratios showing limited sensitivity to annotated variants. For the tasks we study, these results suggest that current NLP-style pretraining strategies provide modest, tokenizer-gated improvements over strong random baselines and motivate more biologically informed tokenization and variant-aware objectives. Our code is available at github.com/m42-health/gfm-random-eval.

genomics↗

bioRxiv2022

Development of an antibody fused with an antimicrobial peptide targeting Pseudomonas aeruginosa: a new approach to prevent and treat bacterial infections

The increase of emerging drug resistant Gram-negative bacterial infections is of global concern. In addition, there is growing recognition that compromising the microbiota, through the use of broad spectrum antibiotics, may affect patient health in the long term. Therefore, there is the need to develop new -cidal strategies to combat Gram-negative infections that would consider these specific issues. In this study, we report and characterize one such approach, the antibody-drug conjugates (ADCs) that combine (i) targeting a specific pathogenic organism through a monoclonal antibody with (ii) the high killing activity of antimicrobial peptides. We focused on a major pathogenic Gram-negative bacterium associated with antibacterial resistance: Pseudomonas aeruginosa and designed an ADC by fusing an antimicrobial peptide at the C-terminal end of the VH and/or VL-chain of a monoclonal antibody, VSX, that targets the core of P. aeruginosa lipopolysaccharide (LPS). This ADC demonstrated appropriately minimal levels of toxicity to mammalian cells and rapidly kills P. aeruginosa strains through several mechanisms while protecting mice from P. aeruginosa lung infection when administered therapeutically. Furthermore, we found that the ADC was synergistic with several classes of antibiotics. This approach described in this study may result in a widely useful strategy to target specific pathogenic microorganisms without augmenting further antibiotic resistance. Author SummaryThe increasing of emerging drug resistant bacterial infections is a worldwide issue and infections caused by antibiotic resistant Gram-negative pathogens are particularly concerning. In addition, there is now growing recognition that disruption of the microbiota, through the use of broad spectrum antibiotics, may affect patient health in the long term. Therefore, there is the need to develop new -cidal strategies to combat Gram-negative infections while preserving the microbiota and also avoid enhancement of antibiotic resistance. We report and characterize here one such approach by using a specific monoclonal antibody associated with the potent killing activity of antimicrobial peptides in the form of an antibody-drug conjugate (ADC). The selected pathogenic bacterium was Pseudomonas aeruginosa, that presents numerous markers for both innate and acquired antibiotic resistance. The ADC lacked significant cytotoxicity against mammalian cells and was shown to be effective both in vitro and in vivo against P. aeruginosa.

microbiology↗

bioRxiv2021

Microbial induced calcite precipitation can consolidate martian and lunar regolith simulants

We demonstrate that Microbial Induced Calcite Precipitation (MICP) can be utilized for creation of consolidates of Martian Simulant Soil (MSS) and Lunar Simulant Soil (LSS) in the form of a brick. A urease producer bacteria, Sporosarcina pasteurii, was used to induce the MICP process for the both simulant soils. An admixture of guar gum as an organic polymer and NiCl2, as bio-catalyst to enhance urease activity, was introduced to increase the compressive strength of the biologically grown bricks. A casting method was utilized for a slurry consisting of the appropriate simulant soil and microbe; the slurry over a few days consolidated in the form of a brick of the desired shape. In case of MSS, maximum strength of 3.3 MPa was obtained with 10mM NiCl2 and 1% guar gum supplementation whereas in case of LSS maximum strength of 5.65 MPa was obtained with 1% guar gum supplementation and 10mM NiCl2. MICP mediated consolidation of the simulant soil was confirmed with field emission scanning electron microscopy (FESEM) and X-ray diffraction (XRD). Our work demonstrates a biological approach with an explicit casting method towards manufacturing of consolidated structures using extra-terrestrial regolith simulant; this is a promising route for in situ development of structural elements on the extra-terrestrial habitats.

bioengineering↗

bioRxiv2020

Inflammation Drives Alternative First Exon usage to Regulate Immune Genes including a Novel Iron Regulated Isoform of Aim2

Determining the layers of gene regulation within the innate immune response is critical to our understanding of the cellular responses to infection and dysregulation in disease. We identified a conserved mechanism of gene regulation in human and mouse via changes in alternative first exon (AFE) usage following inflammation, resulting in changes to isoform usage. Of these AFE events, we identified 50 unannotated transcription start sites (TSS) in mice using Oxford Nanopore native RNA sequencing, one of which is the cytosolic receptor for dsDNA and known inflammatory inducible gene, Aim2. We show that this unannotated AFE isoform of Aim2 is the predominant isoform transcribed during inflammation and contains an iron-responsive element in its 5′UTR enabling mRNA translation to be regulated by iron levels. This work highlights the importance of examining alternative isoform changes and translational regulation in the innate immune response and uncovers novel regulatory mechanisms of Aim2.Summary Sentence Alternative first exon usage was the major splicing event observed in macrophages during inflammation, which resulted in the elucidation of a novel isoform and iron mediated regulatory mechanism of the protein coding gene, Aim2.Competing Interest StatementThe authors have declared no competing interest.View Full Text

immunology↗

bioRxiv2020

BACTERIAL GROWTH INDUCED BIOCEMENTATION TECHNOLOGY, 'SPACE BRICK' - A PROPOSAL FOR EXPERIMENT AT MICROGRAVITY AND PLANETARY ENVIRONMENTS

We present results of our investigation of microbial induced calcite precipitation for manufacturing space bricks and a proposal for study of this activity in low-earth orbit (LEO). Sporosarcina pasteurii, a urease producing bacterial strain was used to consolidate lunar simulant soil (LSS) in the form of a brick with non-trivial strength properties. Potential of a naturally occurring polymer namely, guar gum, as an additive was investigated for enhancement in compressive strength of bio-consolidated samples. Experimental results of bio-brick exhibited an approximate 10-fold increase in compressive strength with guar gum supplementation in soil. We present results of microstructural analysis of the space bricks and also propose a payload design for related experiments in LEO.

bioengineering↗

bioRxiv2019

Toward Broad Spectrum DHFR inhibitors Targeting Trimethoprim Resistant Enzymes Identified in Clinical Isolates of Methicillin-Resistant Staphylococcus aureus

The spread of plasmid borne resistance enzymes in clinical Staphylococcus aureus isolates is rendering trimethoprim and iclaprim, both inhibitors of dihydrofolate reductase (DHFR), ineffective. Continued exploitation of these targets will require compounds that can broadly inhibit these resistance-confering isoforms. Using a structure-based approach, we have developed a novel class of ionized non-classical antifolates (INCAs) that capture the molecular interactions that have been exclusive to classical antifolates. These modifications allow for a greatly expanded spectrum of activity across these pathogenic DHFR isoforms, while maintaining the ability to penetrate the bacterial cell wall. Using biochemical, structural and computational methods, we are able to optimize these inhibitors to the conserved active sites of the endogenous and trimethoprim resistant DHFR enzymes. Here, we report a series of INCA compounds that exhibit low nanomolar enzymatic activity and potent cellular activity with human selectivity against a panel of clinically relevant TMPR MRSA isolates.

biochemistry↗

bioRxiv2018

Structural and Biochemical Studies of Dihydrofolate Reductase from Streptococcus pyogenes as a Target for Antifolate Antibiotics

Streptococcus pyogenes, a beta-hemolytic bacterium, causes a wide spectrum of infections in human including pharyngitis, tonsillitis, scarlet fever, rheumatic fever, and necrotizing fasciitis. Streptococcal infections can also exist as co-infection with methicillin resistant Staphylococcus aureus (MRSA). Trimethoprim-sulfamethoxazole (TMP-SMX) combination has been used for treatment of S. pyogenes and MRSA co-infection. However, resistance to TMP, an inhibitor of dihydrofolate reductase enzyme (DHFR), has challenged the efficacy of TMP-SMX combination. We explored the activity of a series of novel DHFR inhibitors against S. pyogenes. This study identified potent inhibitors of DHFR enzyme from S. pyogenes with excellent inhibitory activity against the growth of the live bacteria. We determined, for the first time, the crystal structure of S. pyogenes DHFR which provides structural insights into design and development of antifolate agents against this global pathogen.

microbiology↗