Sébastien Lemieux

PHATE-Derived Ancestry Coordinates Capture Continuous Population Structure in Human Genomic Data

This paper shows PHATE can better capture continuous population structure in genomic data than PCA or UMAP by tuning scale and balancing populations.

A biological-response compound representation allows chemical perturbation prediction across cell lines

The paper introduces BioPert, a framework that uses gene expression responses to predict chemical perturbations across cell lines, achieving higher accuracy than chemical structure-based methods.

A Transparent and Generalizable Deep Learning Framework for Genomic Ancestry Prediction

This paper develops a deep learning framework for genomic ancestry prediction that is generalizable across diverse populations and interpretable using attribution techniques.

STING dampens the unfolded protein response to enable the presentation of self-antigens on MHC-I during inflammation.

The paper shows STING dampens the unfolded protein response to allow self-antigen presentation on MHC-I during inflammation, linking it to immune regulation in Parkinson's disease.

EZH2 Inhibition Induces a Metabolic Stress Response Sensitizing TNBC to Glutaminase Targeting

This study shows that EZH2 inhibition in TNBC causes metabolic stress, making cells more sensitive to glutaminase inhibitors.

Predictive Gene Discovery with EPCY: A Density-Based Alternative to DE analysis

EPCY ranks genes by predictive power using cross-validated classifiers and density estimation, outperforming DEG-based methods in selecting robust candidate genes.

DEK::NUP214 acts as an XPO1-dependent transcriptional activator of essential leukemia genes

This study identifies DEK::NUP214 as an XPO1-dependent transcriptional activator of leukemia genes, showing selective response to XPO1 inhibitors in t(6;9) AML cases.

Replication of a GWAS signal near HLA-DQA2 with acute myeloid leukemia using a disease-only cohort and external population-based controls

The paper replicates a GWAS signal near HLA-DQA2 associated with AML risk using a disease-only cohort and external population-based controls, finding a stronger effect in patients with NPM1 mutations.

Pepid: a Highly Modifiable, Bioinformatics-Oriented Peptide Search Engine

Pepid is a flexible peptide search engine designed for bioinformatics research, enabling modifications and integration of deep learning methods while maintaining competitive performance.

Mining Mass Spectra for Peptide Facts

The paper improves peptide identification by combining database-driven and de novo approaches through a deep learning-based length prediction model.

Towards Computing Attributions for DimensionalityReduction Techniques

The paper computes local feature attributions for t-SNE using gradient-based methods and validates them on synthetic and biological datasets.

Baseline Acute Myeloid Leukemia Prognosis Models using Transcriptomic and Clinical Profiles by Studying the Impacts of Dimensionality Reductions and Gene Signatures on Cox-Proportional Hazard

This paper evaluates how dimensionality reduction techniques and gene signatures affect Cox-Proportional-Hazard models for AML prognosis using the Leucegene dataset.

CAMAP: Artificial neural networks unveil the role of codon arrangement in modulating MHC-I peptides presentation

The paper introduces CAMAP, an artificial neural network that predicts MHC-I peptide presentation by analyzing codon arrangement in mRNA sequences, improving prediction accuracy beyond traditional methods.

Target variant detection in leukemia using unaligned RNA-Seq reads

This paper introduces km, a method for detecting targeted mutations in AML using unaligned RNA-Seq reads, showing it is fast, accurate, and limited by sequencing depth.