Pratik Katte

Pratik Katte

@pratikkatte.bsky.social

PhD in computational bio @UCSC • Understanding evolution• building ARGs at scale • http://lorax.ucsc.edu • https://pratikkatte.github.io/

15 followers19 following7 posts

Pratik Katte reposted

alex ioannidis
@alexgioannidis.bsky.social

ARGformer is out in Bioinformatics! doi.org/10.1093/bioi... Just as PCA reduces dimensionality from genotypes, ARGformer can do this now for ARGs...yielding a representation that comes directly from inferred genealogical topologies & at a haplotype level across each individual genome.

doi.orgARGformer: learning on ancestral recombination graphs with transformersAbstractMotivation. Recent advances in inference of the ancestral recombination graph (ARG), which describes how segments of chromosomes trace back through
Pratik Katte
@pratikkatte.bsky.social

I’ve been reading statistical mechanics for molecular dynamics alongside coalescent theory for evolution, and noticed an idea they share: Sometimes, taking a system toward infinity makes it simpler to understand. Here’s a beautiful example from evolution ↓

View full thread
Pratik Katte
@pratikkatte.bsky.social

Now let the population become enormous, while keeping our sample small. A collision becomes rare. Multiple collisions in the same generation become even rarer. The messy process simplifies to: wait → two merge → wait → two merge → … → one common ancestor.

Pratik Katte
@pratikkatte.bsky.social

This large-population limit gives us Kingman’s coalescent, a foundation of population genetics. That’s what I find beautiful about infinity: the real system needn’t be infinite. Sometimes the limit simply strips away unlikely details and reveals the structure underneath.

Pratik Katte
@pratikkatte.bsky.social

I’ve been reading statistical mechanics for molecular dynamics alongside coalescent theory for evolution, and noticed an idea they share: Sometimes, taking a system toward infinity makes it simpler to understand. Here’s a beautiful example from evolution ↓

Pratik Katte
@pratikkatte.bsky.social

Take DNA from a few people today and trace those genetic lineages backward. Each came from an ancestral copy in the previous generation. Eventually, two lineages may trace to the same ancestral copy. They coalesce. But how often does this happen?

Pratik Katte
@pratikkatte.bsky.social

The Wright–Fisher model gives a neat picture: DNA lineages = balls Possible ancestors = boxes Throw the balls into boxes. Two in one box → they coalesce. Several collisions → several lineages merge at once. For a finite population, this can get complicated.

Pratik Katte
@pratikkatte.bsky.social

Excited to be presenting my poster on Lorax at #PEQG26! Lorax is a GPU-accelerated, web-native tool for real-time interactive visualization of Ancestral Recombination Graphs. If you're at Asilomar this week, come check it out. Would love to chat about scalable ARG exploration and get your feedback

lorax.ucsc.eduLorax

Pratik Katte reposted

Jeffrey Ross-Ibarra
@jrossibarra.bsky.social

Recently we're working with SNPs from whole genome assemblies to estimate ARGs. It's a pain to go from alignment files to vcf, keeping track of masked and invariant sites. So we wrote a snakemake/SLURM pipeline. Hope it's useful to others, and don't hesitate to post issues if there are problems!

github.comGitHub - RILAB/argprep: Snakemake pipeline for generating SINGER input files from whole genome alignment .maf files.Snakemake pipeline for generating SINGER input files from whole genome alignment .maf files. - RILAB/argprep

Pratik Katte reposted

Molecular Biology and Evolution
@molbioevol.bsky.social

Using ~8M SARS-CoV-2 genomes, a new phylogenetic algorithm by @russcd.bsky.social detects 10 shifts in mutation patterns, some of them linked to changes in viral polymerase. This provides new perspectives on the evolution of mutational processes. 🔗 doi.org/10.1093/molb... #molbio #evobio #covid

doi.orgA Phylogenetic Method Identifies Candidate Drivers of the Evolution of the SARS-CoV-2 Mutation SpectrumAbstract. The molecular processes that generate new mutations evolve, but the causal mechanisms are largely unknown. In particular, the relative rates of m

You are up to date.