Johann A. Gagnon-Bartsch

This paper introduces Wasserstein Geodesic Transformations to improve analysis of high-throughput cell imaging data by addressing long-tailed distributions and outliers.

This paper identifies major sources of unwanted variation in cancer RNA-seq data and proposes a novel method to remove them using pseudo-replicates of pseudo-samples.

The paper develops a method to remove technical artifacts from high-throughput imaging assays using systematic replication, improving biological signal clarity.

This study evaluates deconvolution methods for human brain transcriptome data, finding that cell-type signatures impact accuracy more than methods, and partial methods outperform complete ones.

This paper introduces a hybrid scale approach for cell-type deconvolution that improves biological plausibility and statistical efficiency compared to log or linear scales.

The paper identifies stably expressed genes in scRNA-seq by analyzing genes associated with cellular structures, finding ribosomal genes to be highly stable in total RNA content.

The paper introduces a new normalization method, RUV-III, for Nanostring nCounter gene expression data and evaluates its effectiveness across multiple datasets.

The paper investigates how preprocessing transformations improve the discovery of biological and technical effects in microenvironment microarray data.

scMerge integrates multiple scRNA-Seq datasets by removing unwanted variation through stably expressed genes and pseudo-replicates, enhancing biological discovery.

dtangle deconvolves cell-type proportions from bulk RNA-seq data, showing competitiveness with existing methods and utility in analyzing immune responses to Lyme disease.