Leopold Parts

The paper uses single-cell transcriptomics and phenotypic analysis of gene knockouts in hiPSC-derived neurons and astrocytes to identify cell-type specific and shared mechanisms of Alzheimer's disease risk genes.

The paper builds a spatiotemporal map of human pilosebaceous unit development using single-cell multiomics and machine learning, showing how epithelial-mesenchymal interactions drive organogenesis and how sebaceous gland programs are disrupted in tumors.

The paper identifies suppressor variants that alter gene essentiality across yeast strains, revealing context-dependent genetic mechanisms.

This paper develops a pooled paired prime editing screen to generate large-scale deletions across the genome, identifying non-coding regions with fitness effects and demonstrating the feasibility of genome-wide deletion perturbation.

PARPAL is a web database that analyzes protein redistribution in yeast after paralog deletion using high-content screening and deep learning, providing insights into gene retention and evolution.

This paper develops a CRISPR-based method to generate large-scale structural genomic rearrangements in human cells, revealing selection pressures and gene expression impacts of these changes.

The paper uses enhancer scrambling to study a super-enhancer of the OTX2 gene, finding that a single DNase I hypersensitive site predominantly controls its expression.

The paper presents a genome-scale CRISPRi perturbation map in human pluripotent stem cells using single-cell RNA sequencing to study trans gene regulation across diverse genetic backgrounds.

The paper investigates how DNA repair genes influence Cas9-induced mutations by analyzing mutational outcomes across 2,838 synthetic targets in mouse cells, revealing gene-specific roles and improving predictive models.

The paper identifies misexpression of inactive genes in whole blood and links it to nearby rare structural variants using RNA and genome sequencing data.

This study uses single-cell imaging to analyze protein dynamics in response to paralog deletion, revealing mechanisms of gene retention and functional divergence.

ArtSeg uses a deep learning model to automatically segment and remove artifacts from brightfield microscopy images using image-level labels, improving downstream analysis without requiring pixel-level annotations.

This paper identifies position-specific sequence determinants of base editing efficiency and develops a machine learning model to predict editing outcomes for adenine and cytosine editors.

The paper predicts insertion efficiency of short sequences into the genome using prime editing by analyzing sequence features and developing a machine learning model.

The paper measures how natural genetic variation can suppress mutations in essential yeast genes, finding that 35% of tested genes can be suppressed by modifier genes in different backgrounds.

The paper designs a smaller genome-wide CRISPR-Cas9 library with reduced size while maintaining sensitivity and specificity, enhancing its applicability in complex models.

The paper uses CRISPRi/a to screen gene functions in yeast, identifying core signaling pathways and novel interactions involved in heat adaptation and stress response.

This paper introduces a minimized double guide RNA library for CRISPR/Cas9 screens, reducing scale by fivefold while maintaining performance.

This paper demonstrates accurate nucleus segmentation from brightfield images using deep learning, showing that three CNN architectures can achieve this with U-Net performing best.

JACKS improves hit identification in CRISPR/Cas9 screens by jointly analyzing data from the same guide RNA library, reducing required cell numbers by 2.5-fold.

This paper predicts antibiotic resistance in E. coli from whole genome sequences using machine learning, achieving high accuracy without prior knowledge of resistance mechanisms.

The paper predicts Cas9-induced mutation outcomes based on surrounding DNA sequence using data from over 1 billion mutational outcomes.

The paper identifies an ISRIB-binding site in eIF2B using cryo-EM and CRISPR/Cas9, showing its role in regulating translation through an allosteric mechanism.

The paper predicts quantitative traits in yeast with high accuracy using family data and genetic variants, showing that close relatives improve predictions more than large cohorts.

This paper introduces Phased Outbred Lines (POLs) to decompose trait variation in diploid organisms, achieving high heritability and precise estimation of genetic components in yeast.