David B. Blumenthal

NetMap jointly infers cell-level gene regulatory networks and clusters cells based on shared regulatory programs, outperforming existing methods on synthetic data and reliably detecting regulons on real data.

scXMatch quantifies distribution shifts between single-cell data from different conditions using a non-parametric graph-based test, showing robust performance across gene expression, chromatin accessibility, and morphology data.

The paper introduces MaSkel and napari-MaSkel, tools for fast skeletonization and graph-based feature extraction from segmentation masks, using an optimized Lee94 thinning algorithm.

POME provides graph-based embeddings for partially observed mixed-type biomedical data, achieving state-of-the-art imputation and enabling downstream analyses like subgroup discovery and predictive modeling.

Inferring tested-but-undetected pairs from prey lists of 5,530 published interaction screens yields 97 million tested human protein pairs and 18,982,508 non-interactions (≥3 tests) for calibrating screens and predictors.

This study benchmarks graph neural networks for cancer subtyping using methylation and gene expression data, finding that ChebNet performs best and simple readout layers improve performance.

SignifiKANTE computes efficient P-values for gene regulatory networks by clustering target genes to reduce computational cost.

FLAG-X provides a Python package with hybrid workflows for automated flow cytometry gating, integrating manual and machine learning methods to improve consistency and efficiency in clinical settings.

This study shows TNF-α disrupts the malate-aspartate shuttle in iPSC-derived enteric neural cells from PD patients, causing metabolic shifts and mitochondrial dysfunction, which can be reversed by targeting glutamate metabolism.

GNExT is a web-based platform that integrates GWAS data with systems medicine approaches to identify potential drug repurposing candidates and enable large-scale network medicine investigations.

This study identifies how Leishmania donovani hijacks host ferritin to create an iron reservoir, protecting against oxidative stress and ferroptosis.

AlterNet infers transcript-level gene regulatory networks by considering alternative splicing, revealing novel regulatory interactions not detectable at the gene-level in heart tissue data.

This study quantifies morphodynamic features of peritoneal resident tissue macrophages in vivo to distinguish functional states and assess tissue age.

This study shows that sequence-based PPI prediction models using ESM-2 embeddings plateau at 0.65 accuracy, suggesting structural information may be needed for better performance.

This paper introduces SwitchTFI, a method to identify transcription factors driving cell differentiation using single-cell RNA sequencing data, and shows it outperforms existing methods in functional coherence.

TRIPS identifies disease-associated transcription factors by integrating differential expression and network perturbation analysis in case-control data.

This study shows that power-law distributions in PPI networks can arise from study and technical biases, not biological properties.

DysRegNet infers patient-specific gene-regulatory networks from bulk RNA-seq data, accounting for confounders and scaling to large sample sizes, showing age-specific bias in breast cancer.

The paper analyzes spatial protein abundance data from skin tissue samples to identify characteristic spatial organization patterns in CTCL compared to AD and PSO.

This paper predicts 5-year progression-free survival in melanoma using spatial proteomics and few-shot learning, achieving over 90% accuracy.

The paper introduces NeEDL, a network medicine-based method for detecting higher-order epistatic interactions in complex diseases, showing improved statistical significance and applicability across eight diseases.

The paper evaluates deep learning models for protein-protein interaction prediction, finding they rely on data leakage and sequence similarity, and advocates for baseline comparisons.

DataSAIL splits biological datasets to minimize information leakage by reducing similarity between splits, improving model generalization assessment.

TF-Prioritizer identifies condition-specific transcription factors using multi-modal genomic data and generates an interactive web report.

BoostDiff identifies differential gene regulatory networks between disease and healthy states, outperforming existing methods on simulated and real transcriptomics data.

BiCoN identifies patient subgroups and functional gene subnetworks by integrating network constraints with biclustering of omics data.