Jeffrey A. Bilmes

ChromaFold predicts 3D chromatin contact maps from scATAC-seq data, outperforming bulk-based methods and enabling cell-type-specific interaction analysis.

Epiphany predicts cell-type-specific Hi-C contact maps from epigenomic signals, showing robust performance across cell types and species.

MS1Connect measures similarity between mass spectrometry runs using MS1 scans, outperforming baselines in species prediction and correlating with MS2 scan similarities.

This paper introduces Synmatch, a method to align cells across single-cell modalities by matching neighborhood structures, showing it works well on small datasets and scales to large ones.

This paper introduces a deep learning model called GLEAMS to embed mass spectra into a low-dimensional space for efficient clustering and analysis of large-scale proteomics data.

DIAmeter is a library-free search engine for analyzing DIA mass spectrometry data using peptide sequences, supporting various isolation windows and post-translational modifications.

This paper identifies a bias in machine learning models that predict genomic features across cell types when trained on data from the same genomic loci across multiple cell types.

The paper demonstrates using submodular optimization to create efficient, high-quality sketches of single-cell RNA-seq data.

The paper develops a method to prioritize genomic and epigenomic experiments using imputed data to maximize information gain, and introduces a software framework called Kiwano.

The paper introduces Avocado, a deep tensor factorization method that learns a latent representation of the human epigenome, improving imputation and performance on genomics tasks.

The paper uses Avocado to impute epigenomic and transcriptomic data across 400 human biosamples and 84 assays, improving protein binding predictions.

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 regression model called Siren to improve peptide detection in DIA mass spectra by accounting for isotope peak interference.

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.

This paper introduces THEMIS, a method for analyzing tumor heterogeneity across multiple biopsies, showing improved accuracy over existing methods and applications in tracking cancer progression and treatment resistance.

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