Maxwell W. Libbrecht

This study evaluates if ensembling improves feature attributions from sequence-to-activity models, finding that it enhances downstream applications like motif identification and variant prediction.

The paper introduces a pan-cell type chromatin state annotation that integrates all IHEC epigenomes using the epigenome-ssm method, avoiding cell type-specific annotations.

GAME provides a standardized framework for benchmarking genomics models across tasks, enabling consistent evaluation and community contributions.

CANDI improves epigenome imputation by predicting raw counts, handling covariates, and providing uncertainty measures using a self-supervised Transformer model.

VSS-Hi-C stabilizes variance in Hi-C data to improve visualization and analysis by transforming read counts into unit variance signals.

The paper evaluates the reproducibility of SAGA chromatin state annotations and introduces SAGAconf to assign confidence scores to these annotations.

The paper presents ENCODE4 Segway annotations for 234 human cell types, improving sensitivity in capturing regulatory patterns relevant to disease and traits.

The paper introduces an integrative method for chromatin domain annotation using Hi-C data and functional genomic signals without a pairwise prior, achieving better resolution of subcompartments.

Hi-C-LSTM uses a recurrent neural network to learn low-dimensional representations of Hi-C data, enabling the identification of genomic elements that define chromatin conformation.

SigTools provides visualization tools for genomic signals to aid in data exploration and interpretation, with case studies on histone modifications and novel chromatin features.

The paper introduces a method to identify driver cell populations in flow cytometry data using a new score called SpecEnr and provides an R package for implementation.

This paper introduces an interpretable classification method for predicting drug resistance in M. tuberculosis, achieving high accuracy and flexibility.

The paper introduces VSS, a method to stabilize variance in sequencing-based genomic signals, improving downstream applications like SAGA.

This paper introduces Epi-LSTM, a deep LSTM autoencoder that creates a pan-celltype epigenomic representation, outperforming existing methods in most cell types and providing smoother genomic representations.

This study uses machine learning on whole-genome sequencing data from multiple geographic regions to predict drug resistance in Mycobacterium tuberculosis, finding that performance varies across datasets due to geographical heterogeneity.

This paper introduces a semi-supervised method for genome annotation that improves accuracy by incorporating expert annotations and evaluates it using a new precision and recall metric.

The paper develops a fully automated method to annotate 164 human cell types using 1,615 genomics datasets, creating a unified encyclopedia of functional DNA elements with a new importance measure.

The paper introduces a method to annotate the human epigenome with continuous chromatin state features, improving downstream applications like gene and enhancer identification.

Rambutan, a convolutional network, calls high-confidence 1 kb Hi-C contacts from DNA sequence plus DNaseI signal across six cell types; the authors flag an error in contact calling.

PREDICTD imputes missing epigenomics experiments using tensor decomposition, improving accuracy over existing methods and enabling analysis of enhancer biology in non-coding human regions.

Segway 2.0 improves genome annotation by using Gaussian mixture models and minibatch training for more accurate signal pattern discovery.