Thomas Von Yu

Community assessment of methods to deconvolve cellular composition from bulk gene expression

The paper evaluates deconvolution methods for inferring immune infiltration from bulk gene expression using a DREAM Challenge, finding that some community-contributed methods, especially deep learning-based ones, perform well.

Insights from the IronTract challenge: optimal methods for mapping brain pathways from multi-shell diffusion MRI

The paper evaluates tractography methods using a macaque dataset to determine optimal approaches for mapping brain pathways from multi-shell diffusion MRI.

Cell-to-cell and type-to-type heterogeneity of signaling networks: Insights from the crowd

The paper organizes a DREAM challenge to benchmark predictive models of single-cell signaling using mass cytometry data from breast cancer cell lines.

Tractography dissection variability: what happens when 42 groups dissect 14 white matter bundles on the same dataset?

This study evaluates variability in white matter bundle segmentation across 42 groups using the same dataset, finding protocol variability exceeds other sources of inconsistency.

Crowdsourcing assessment of maternal blood multi-omics for predicting gestational age and preterm birth

The paper develops predictive models for preterm birth using maternal blood multi-omics data, finding transcriptomic and proteomic markers can predict gestational age and spontaneous preterm birth with high accuracy.

Predicting cellular position in the Drosophila embryo from Single-Cell Transcriptomics data

The paper predicts cellular positions in the Drosophila embryo using scRNA-seq data by benchmarking methods against spatially annotated genes.

DreamAI: algorithm for the imputation of proteomics data

An ensemble of six imputation methods, built through the NCI-CPTAC Proteogenomics DREAM Challenge, fills missing values in labeled LC-MS/MS proteomics data, helping most for low-abundance proteins.

A cancer pharmacogenomic screen powering crowd-sourced advancement of drug combination 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.

A community-based collaboration to build prediction models for short-term discontinuation of docetaxel in metastatic castration-resistant prostate cancer

This paper presents a crowdsourced challenge to predict early discontinuation of docetaxel in mCRPC patients using clinical data from multiple trials, finding that ensemble models outperform individual models.