01
Computational genomics
Genomic sequence, variation, transcriptomics, biological information and the computational analysis of genome-scale data.
Bioinformatics · Computational Biology
Sarvesh
Galgale
MSc Bioinformatics student at REVA University. I work across genomics, machine learning and biological data, with a growing interest in what computational models can actually tell us about biology.
Explore my work ↓
01 / About
I like figuring things out.
I started out studying genetics, microbiology and biochemistry, and my move towards bioinformatics happened gradually. I found myself increasingly interested in what happened when I could take biological data, work with it computationally, find patterns and test questions for myself.
That is still a large part of what I enjoy about research. I like following an unexpected result, writing code to test an idea, or building a workflow that makes a biological question easier to investigate.
Right now, I am moving towards computational biology around genomics and machine learning. I am especially interested in what computational results actually mean biologically, but I also like having room to follow different questions and see where they lead.
02 / Research
Questions I keep coming back to.
My interests are fairly broad, but most of the questions I keep returning to involve understanding what we can learn from biological data, and how far computational models can take us in doing that.
02
Machine learning for biology
Foundation models, representation learning and deep learning, with an interest in understanding what biological models learn rather than treating prediction alone as the endpoint.
03
Biological systems & interpretation
Proteins, microbial systems, disease and computational questions where different kinds of evidence need to be brought together to make biological sense of a result.
03 / Projects
Research I have been working on.
01
Manuscript in preparation
Multiscale superinformation reveals structured organization of genomic sequence features
I looked at how entropy and superinformation vary across genomic features and seven model organisms, asking whether characteristic sequence blocks reveal structure that differs across feature types.
02
In progress
LGMD Foundation Model Benchmark
A benchmark of biological foundation models for representing LGMD-associated SNVs, using six models and four downstream classifiers across 6,962 variants from 10 genes. One protein foundation model is included as a cross-domain comparator.
6 × 4
models × classifiers
03
Completed
AMR & detoxification-associated hypothetical proteins
I built a deep learning framework for poorly characterized bacterial proteins, then followed its predictions with explainability, structural analysis and functional evidence.
Prediction→Explanation→Biological evidence
Project review ongoing.
Integrated Microbiome Analysis of the Godavari River Ecosystem
The first phase used three water samples collected upstream, midstream and downstream along the Godavari River in Nashik, Maharashtra, generating original 16S rRNA amplicon data. Taxonomic profiling and PICRUSt2-inferred functional profiles were used to examine environmental-stress and pollutant-associated signatures.
River sampling→16S sequencing→Microbiome profiling→PICRUSt2→Interpretation
Expanding to riverbed soil, plants and fish gut microbiomes.
04 / Workflow & Pipelines
Reproducible workflows I've built around biological questions.
01
Released · v1.0.0
RNA-seq Pipeline
A Nextflow workflow for RNA-seq preprocessing, alignment, quantification and downstream analysis, with optional fusion-detection and novel-transcript branches.
FASTQ→QC→STAR→Counts→DESeq2
Arriba fusion detection
StringTie / CPC2 transcript discovery
02
Released · v1.0.0
GSEA Analysis
A reproducible enrichment workflow that accepts either raw gene counts or an already ranked gene list.
Raw counts→DESeq2→Ranking
Pre-ranked genes
↓
fgsea → results → plots
03
Validation in progress
v1.0.0-rc1 · Manuscript in preparation
WGS Analysis
A human short-read WGS workflow for germline and matched tumor-normal analysis, with multiple callers, benchmarking, concordance, SV/CNV analysis, annotation and integrated reporting.
FASTQ → caller-ready BAM
Germline
Tumor-normal somatic
BenchmarkingConcordanceSV / CNVReporting
In development
mtDNA Analysis
ctDNA Analysis
Single-nucleus RNA-seq
05 / Publications & Research Output
Papers, manuscripts and posters.
Publication
Molecular Docking and Dynamic Simulation-Based Screening Identifies Inhibitors of Targeted SARS-CoV-2 3CLPRO and Human ACE2
View DOI ↗Manuscripts in preparation
Multiscale superinformation reveals structured organization of genomic sequence features
Why Machine Learning in Biology Needs Models, Mechanisms, and Meaning
WGS pipeline manuscript
Poster presentations
Superinformation Thresholds for Annotating Sequence Ontology Features
Genomics India Conference 2025 · IISc Bangalore
DNA Language Model for Rare Disease Gene and Variant Identification
REDRESS 2025 · Tata Institute of Genetics & Society
06 / Other Work
Things I make because I want to see where the idea goes.
Small tools, experiments and ideas I build around things I find interesting.
01
Ongoing experiment
DNAFreak
DNAFreak really just came from me enjoying the weird side of biology. I wanted somewhere I could type in whatever I was curious about and see what strange connection came out of it. I am still having fun with it, so I will probably keep adding new things as I go.
Visit DNAFreak ↗07 / Academic Journey
How I got here.
2020—2023
BSc Genetics, Microbiology & Biochemistry
AIMS Institutes · Bangalore University
Where it started. A broad grounding in genetics, microbiology and biochemistry, and where my interest in biology really began.
2024—2025
One League Fellowship · PGDM
Stanford Engineering · Entrepreneurship & Product Innovation
I learned about product thinking, innovation and building for people, and started thinking about what those ideas could look like in biotech and research. I still find myself thinking about the design principles we unknowingly use in research, and what else we could bring in to make research more human-centred.
2025—2027
MSc Bioinformatics
REVA University
A move from studying biology to asking biological questions through computation. I have been learning to work with biological data, think about biology in systems, and explore where genomics, machine learning and computational biology come together.
Biology→Product & Design thinking→Computational biology
08 / Beyond Research
There is more to building than code.
Outside research, I have a soft spot for entrepreneurship and building things. I like thinking about why some ideas work, how they become useful, and everything that happens between having an idea and actually getting it into people's hands: design, economics, policy, geography, scale, distribution and all the messy parts in between.
That way of thinking sometimes finds its way back into how I approach research and scientific tools.
09 / Contact
Let's connect.
If you want to talk about computational biology, research, workflows, or something interesting you are building, feel free to reach out.