Bence Szalai

The paper presents an ensemble model that improves predictions of perceptual similarity between odor mixtures using a curated dataset and collaborative challenge results.

RIDDEN predicts receptor activity from receptor-regulated gene expression, outperforming methods that use ligand-receptor co-expression, and identifies mechanistic biomarkers in cancer patients.

This study benchmarks a foundational cell model for predicting post-perturbation RNAseq data and finds simpler models outperform it.

The paper introduces the EFFECT benchmark suite to evaluate cancer sensitivity prediction models by addressing data biases and improving benchmarking accuracy.

The paper develops a method to screen drug combinations using single-cell transcriptome analysis, identifying synergistic and antagonistic pairs in K562 cells.

The paper compares 15 resources and 6 methods for inferring cell-cell communication from single-cell RNA data, finding significant variation in predicted interactions and biological interpretations.

The paper identifies plasma membrane cholesterol depletion as a key mechanism of antiviral drug activity against SARS-CoV-2 through gene expression analysis and functional experiments.

The paper develops patient-specific Boolean models of prostate cancer signaling networks to identify actionable targets for personalized treatment, finding dose-dependent effects of inhibitors on key targets.

The paper investigates why pathway analysis methods perform better than expected, finding that their effectiveness stems from overlaps with transcription factor regulons rather than accurate pathway activity inference.

The paper benchmarks bulk functional genomics tools on scRNA-seq data, showing comparable performance to bulk data and suggesting their applicability over dedicated single-cell tools.

The paper tests if human pathway and TF footprints can be applied to mouse data, showing they recover mouse perturbations and disease associations.

The paper links perturbation transcriptomics data with cell viability phenotypes to identify cell death and proliferation signatures, improving drug sensitivity prediction.

The paper presents a large-scale cancer pharmacogenomic screen with 11,576 experiments on 910 drug combinations across 85 cancer cell lines, enabling crowd-sourced prediction of synergistic drug pairs and biomarkers.

The paper demonstrates that machine learning can accurately predict odor perception from molecular features, enabling reverse-engineering of smell.