Rojalin Pradhan

Postdoctoral research associate in @lab_ogden, @StJudeResearch

[Ralser Lab @RalserLab]

When #AlphaFold2 came out, we wondered: Could it help us understand how enzymes evolve? So we tried it—and it worked! https://t.co/rtGbSpeAvk

[Bedzhov lab / Embryonic Self-Organization @emselor]

Cellular dynamics in the dormant embryo
(highlights of Rui Chen's work on embryo dormancy)

https://t.co/CwaOncKO3A

[Bo Wang @BoWang87]

How can we make genomic foundation models actually useful to biology?! Teach them to REASON!!

🧬 Excited to share BioReason - the first model to successfully integrate DNA foundation models (eg, Evo 2) with LLMs (eg, Qwen3) for biological reasoning!

🔬 What we built:
• Novel multimodal architecture that lets LLMs directly process genomic sequences as input
• Trained with supervised fine-tuning + GRPO reinforcement learning for sophisticated multi-step reasoning
• Generates interpretable step-by-step biological reasoning traces from genomic data

📊 Benchmarks & Performance:
• KEGG pathway reasoning: 97% accuracy on mechanistic variant-to-disease prediction
• Variant effect prediction: 80-88% accuracy on pathogenic/benign classification
• Evaluated on 1,449 KEGG entries + 86K+ ClinVar variants across coding/non-coding regions
• Consistent 15%+ performance gains over DNA-only or LLM-only baselines

🎯 What this truly means:
This enables AI systems to provide mechanistic biological insights comparable to domain experts - bridging the gap between "black box" DNA foundation models and interpretable scientific reasoning. Could fundamentally accelerate hypothesis generation and biological discovery by making genomic AI transparent and actionable.

🙏 Thanks to our amazing team: Adibvafa Fallahpour (@adibvafa) Andrew Magnuson, Purav Gupta, Shihao Ma, Jack Naimer, Arnav Shah, Haonan Duan, Omar Ibrahim, Hani Goodarzi (@genophoria), Chris J. Maddison (@cjmaddison)

🏛️ Across institutions: University of Toronto(@UofT), Vector Institute (@VectorInst ), University Health Network (@UHN), Arc Institute (@arcinstitute), Cohere (@cohere), Google DeepMind(@GoogleDeepMind)

📄 Paper: https://t.co/EhWprAqZnc
🌐 Website: https://t.co/ShxytF4uU0
💻 Code: https://t.co/vKtLNELPe8
🤗 Datasets: https://t.co/ESVnlPjMpM

#AI #MachineLearning #Genomics #Biology #Research #BioAI

[Rojalin Pradhan @Rojalinpr]

https://t.co/mzguU5lA6W

[PrakashLab @PrakashLab]

(1/n) Excited to announce FrugalScience 2025 (https://t.co/JZyDjXNNB6). In its 5th year - Frugal science is a global community of creators engaged in bringing affordable solutions to the world. Anyone around the world can sign up (for free): form here https://t.co/2c3HfXDNxH

[Rojalin Pradhan @Rojalinpr]

https://t.co/9N9ndKImo4

[eLife - the journal @eLife]

The geometric basis of epithelial convergent extension.
https://t.co/zmhqPjyavT

[Crump Lab @CrumpLab]

The outer ear is a mammalian innovation but where did it come from? In our study in @Nature, @MathiThiru95 and colleagues find that the outer ear arose from modification of an ancestral gill program first originating in marine invertebrates. https://t.co/i05me2SXI4

1/n

[eLife - the journal @eLife]

The geometric basis of epithelial convergent extension.
https://t.co/zmhqPjyavT

[eLife - the journal @eLife]

An automated method to track individual mammalian cells as they progress through the cell cycle, demonstrated by testing drugs on tumour cells.
https://t.co/L2VzjD25y9

[Reto Fiolka @RetoPaul]

Maybe the most amazing microscope paper of the year just dropped: Live 4pi-SIM (a.k.a I5S*):
https://t.co/RJale5ulmK

Having worked with Mats Gustafsson, I can attest that this is both extremely hard, but also extremely cool!

Kudos to the authors!

*https://t.co/ysoGgXIlF7

[Samuel Hume @DrSamuelBHume]

2024's top 10 advances in biology (🧵)

1. Progress in longevity

Blocking the pro-inflammatory cytokine, IL11, extended healthspan and lifespan (in mice!)

Old mice without IL11 look healthier, they're stronger, and they live longer

https://t.co/Y0Raog12xk

[Cell Biology J-Club @cell_club]

PunctaFinder: An algorithm for automated spot detection in fluorescence microscopy images | Molecular Biology of the Cell https://t.co/GdNBSeS7tV

[Alice Ting @aliceyting]

Could one envision a synthetic receptor technology that is fully programmable, able to detect diverse extracellular antigens – both soluble and cell-attached – and convert that recognition into a wide range of intracellular responses, from transgene expression and real-time fluorescence to modulation of innate cell behavior (excitation or inhibition of neurons, induction of cell migration, etc.)?

Today we report in Nature a new technology platform that provides a step in that direction: PAGERs, for Programmable Antigen-gated G protein-coupled Engineered Receptors, convert recognition of extracellular soluble or cell-attached antigens into diverse user-selected responses. PAGERs are based on G-protein coupled receptors (GPCRs), which themselves are not structurally modular, but we were able to build in modular antigen gating by fusing an antagonist peptide to the extracellular N-terminal end, and then gating the antagonist with a fused antigen-binding nanobody.  When antigen binds, it sterically interferes with the antagonist, leading to relief of receptor inhibition. Drug or agonist can then turn on PAGER.

https://t.co/Clt8yjyk2y

[Rojalin Pradhan @Rojalinpr]

https://t.co/1zjAtEoFcr

[Rojalin Pradhan @Rojalinpr]

https://t.co/QWYMEGXm2b

[Rojalin Pradhan @Rojalinpr]

https://t.co/uR0jFecBu8

[Srinivas Research Group @LabSrinivas]

Even our embryos are starting to feel a little festive in the Srinivas lab! We are happy to share our new bioRχiv pre-print – investigating the role of calcium in the early embryo

https://t.co/pd3aDAKq4h

[Rojalin Pradhan @Rojalinpr]

https://t.co/eOXml2O6y7