Philip Fradkin

Co-founded blank.bio & Accepted into Y Combinator

Y Combinator (S25)

Co-founded blank.bio, we are building RNA foundation models, and were accepted into the Y Combinator Summer 2025 batch.

mRNABench: A curated benchmark for mature mRNA property and function prediction

Ruian (Ian) Shi* , Taykhoom Dalal* , Philip Fradkin* , Divya Koyyalagunta , Simran Chhabria , Andrew Jung , Cyrus Tam , Defne Ceyhan , Jessica Lin , Kaitlin U. Laverty , Ilyes Baali , Bo Wang , Quaid Morris

bioRxiv (in review)

We present mRNABench, a comprehensive benchmarking suite for mature mRNA biology that evaluates the representational quality of embeddings from self-supervised nucleotide foundation models. *Denotes co-first authorship.

Orthrus: towards evolutionary and functional RNA foundation models

Philip Fradkin* , Ruian Shi* , Taykhoom Dalal* , Keren Isaev , Brendan J. Frey , Leo J. Lee† , Quaid Morris† , Bo Wang†

bioRxiv (in review)

Orthrus is a Mamba-based mature RNA foundation model pre-trained using a novel self-supervised contrastive learning objective with biological augmentations from splice isoforms and orthologous genes. *Denotes co-first authorship. †Denotes co-supervising authorship.

On the Scalability of GNNs for Molecular Graphs

Maciej Sypetkowski , Frederik Wenkel , Farimah Poursafaei , Nia Dickson , Karush Suri , Philip Fradkin , Dominique Beaini

Advances in Neural Information Processing Systems (NeurIPS 2024)

We analyze the scaling behavior of message-passing networks, graph Transformers, and hybrid architectures on the largest public collection of 2D molecular graphs, demonstrating that GNNs benefit tremendously from increasing scale.

How molecules impact cells: Unlocking contrastive phenomolecular retrieval

Philip Fradkin* , Puria Azadi Moghadam* , Karush Suri , Frederik Wenkel , Ali Bashashati , Maciej Sypetkowski , Dominique Beaini

Advances in Neural Information Processing Systems (NeurIPS 2024)

We propose MolPhenix, a model that learns a joint latent space between molecular structures and cellular morphology, significantly improving zero-shot molecular retrieval for drug discovery. *Denotes co-first authorship.

IsoCLR: Splicing Up Your Predictions with RNA Contrastive Learning

Philip Fradkin* , Ruian Shi* , Bo Wang , Brendan Frey , Leo J. Lee

ICML 2023 Workshop on Computational Biology

Presented at the ICML Workshop on Computational Biology. *Denotes co-first authorship.

Robustness to Adversarial Gradients

Philip Fradkin , Lazar Atanackovic , Michael R. Zhang

ICML 2022 Workshop on Pre-training

We evaluate minima sharpness by taking an adversarial step and measure the change in loss.

Concerto: A graph neural network approach for molecule carcinogenicity prediction

Philip Fradkin , Adamo Young , Lazar Atanackovic , Brendan Frey , Leo J Lee , Bo Wang

Bioinformatics / ISMB 2022

Our publication was accepted to Bioinformatics and I got to present our work at ISMB 2022.

Started PhD

University of Toronto

I started my PhD at University of Toronto with Brendan Frey and Bo Wang!

ATP7B variant c.1934T > G p.Met645Arg causes Wilson disease by promoting exon 6 skipping

Daniele Merico , Carl Spickett , Matthew O’Hara , Boyko Kakaradov , Amit G Deshwar , Philip Fradkin , Shreshth Gandhi , Jiexin Gao , Solomon Grant , Ken Kron , Frank W Schmitges , Zvi Shalev , Mark Sun , Marta Verby , Matthew Cahill , James J Dowling , Johan Fransson , Erno Wienholds , Brendan J Frey

NPJ Genomic Medicine

We demonstrated that a specific ATP7B variant, prevalent in Wilson disease, causes a splicing error leading to loss of function, clarifying its pathogenic mechanism.