Jeffrey T. Leek

Insights into the Datasets, Tools, and Training Needs of the AnVIL Community: 2024

This study surveys the AnVIL community to understand user needs, training preferences, and platform adoption opportunities.

Generative genomics accurately predicts future experimental results

This paper introduces GEM-1, a generative model that accurately predicts future gene expression experiments, achieving performance comparable to real lab experiments.

Genotype prediction of 336,463 samples from public expression data

The paper predicts genotypes for 336,463 samples using RNA-seq data from Recount3, achieving high accuracy and enabling population structure analysis.

Inverting the model of genomics data sharing with the NHGRI Genomic Data Science Analysis, Visualization, and Informatics Lab-space (AnVIL)

The paper describes AnVIL, a cloud-based platform that enables genomic data analysis without data transfer, improving security, scalability, and collaboration.

recount3: summaries and queries for large-scale RNA-seq expression and splicing

recount3 provides uniformly processed RNA-seq data and tools for querying and analyzing large-scale human and mouse expression and splicing data.

Strategies for cellular deconvolution in human brain RNA sequencing data

The paper shows that incorporating cell size parameters improves deconvolution of brain RNA-seq data, but using external cell size estimates can introduce bias.

Functional Annotation of Human Long Non-Coding RNAs via Molecular Phenotyping

The paper functionally annotates human long non-coding RNAs by systematically knocking down 285 lncRNAs and analyzing molecular phenotypes using CAGE sequencing.

Recounting the FANTOM Cage Associated Transcriptome

The paper presents FC-R2, an expanded transcriptome atlas of over 109,000 genes, including lncRNAs, and demonstrates its utility in analyzing gene expression across normal and diseased human samples.

Post-prediction inference

The paper develops a method to correct statistical inference using predicted outcomes from machine learning models, demonstrating its effectiveness in two distinct applications.

Addressing confounding artifacts in reconstruction of gene co-expression networks

This study shows that removing top principal components from gene expression data improves gene co-expression network reconstruction by reducing false discoveries, using GTEx data and simulated examples.

recount-brain: a curated repository of human brain RNA-seq datasets metadata

The paper curates a database of human brain RNA-seq metadata to improve reproducibility and analysis of gene expression data.

A direct approach to estimating false discovery rates conditional on covariates

The paper proposes a regression framework to estimate false discovery rates conditional on covariates and applies it to genome-wide association studies.

Developmental And Genetic Regulation Of The Human Cortex Transcriptome In Schizophrenia

The paper identifies 237 genes associated with schizophrenia using RNA sequencing and eQTL analysis, showing enrichment for developmentally regulated synaptic genes.

RNA-seq transcript quantification from reduced-representation data in recount2

The paper develops a linear model to estimate transcript-level abundances from reduced-representation RNA-seq data in recount2, evaluating it on simulated and real data.

crsra: A package for Cleaning and Analyzing Coursera Research Export Data

The paper introduces the crsra package in R for cleaning and analyzing Coursera research export data, offering tools for data loading, integration, and student behavior analysis.

Improving the value of public RNA-seq expression data by phenotype prediction

The paper predicts missing phenotype annotations for public RNA-seq data using well-annotated datasets, enabling better analysis of large-scale genomic data.