Brendan J. Frey

Orthrus: Towards Evolutionary and Functional RNA Foundation Models

Orthrus is a Mamba-based RNA foundation model pre-trained with biological augmentations and contrastive learning, achieving better mRNA property prediction than existing models with less fine-tuning data.

Enigma: An Efficient Model for Deciphering Regulatory Genomics

Enigma is an efficient regulatory genomics model that achieves competitive performance with state-of-the-art models while significantly reducing computational cost.

FlashRNA: An Efficient Model for Regulatory Genomics

FlashRNA improves computational efficiency for genomic sequence-to-function models while maintaining or enhancing predictive performance across multiple tasks.

Sequence based prediction of cell type specific microRNA binding and mRNA degradation for therapeutic discovery

This paper introduces REPRESS, a deep learning model that predicts cell-type-specific microRNA binding and mRNA degradation from RNA sequences, outperforming existing methods in multiple tasks.

An RNA foundation model enables discovery of disease mechanisms and candidate therapeutics

The paper introduces BigRNA, an RNA foundation model that predicts RNA biology from DNA sequence and identifies therapeutic effects of SBOs on gene expression and splicing.

A graph neural network approach to molecule carcinogenicity prediction

This paper introduces a graph neural network model for predicting molecule carcinogenicity, showing improved performance over existing methods.

ATP7B Variant c.1934T>G p.Met645Arg Causes Wilson Disease by Promoting Exon 6 Skipping

This paper shows that the ATP7B variant c.1934T>G causes exon 6 skipping, leading to loss of function and confirming its pathogenicity in Wilson Disease.

HiCSR: a Hi-C super-resolution framework for producing highly realistic contact maps

HiCSR enhances low-resolution Hi-C data to produce high-resolution contact maps with improved accuracy and biological relevance.

cDeepbind: A context sensitive deep learning model of RNA-protein binding

The paper introduces cDeepbind, a deep learning model that incorporates RNA secondary structure to improve RBP binding predictions, achieving better performance on both in-vitro and in-vivo data.

COSSMO: Predicting Competitive Alternative Splice Site Selection using Deep Learning

COSSMO predicts competitive alternative splice site selection using deep learning, achieving 70% accuracy in identifying the most used splice site and 60% R2 in modeling PSI distribution.

Inference Of The Human Polyadenylation Code

The paper develops a deep learning model to predict tissue-specific polyadenylation site selection using genomic features and 3-end sequencing data.

Allele-specific transcription factor binding as a benchmark for assessing variant impact predictors

The paper benchmarks variant impact predictors using allele-specific binding ChIP-seq data, finding deep learning methods perform best but have room for improvement.

Probabilistic estimation of short sequence expression using RNA-Seq data and the positional bootstrap

The paper introduces the positional bootstrap method to improve accuracy in estimating short sequence expression from RNA-seq data by quantifying uncertainty from non-uniform coverage.