Gregory P. Way

On doubting image quality assessment metrics for microscopy virtual staining

This study evaluates image quality assessment metrics for virtual staining in microscopy, finding they poorly translate to biological contexts and suggest application-specific model evaluation is needed.

Single-cell hit calling in high-content imaging screens with Buscar

Buscar is a method for hit calling in high-content imaging screens that leverages single-cell morphology profiles to quantify perturbation efficacy and specificity.

Scalable data harmonization for single-cell image-based profiling with CytoTable

The paper introduces CytoTable, a software package that harmonizes single-cell image-based profiling data from different tools, enabling reproducible and scalable data integration.

Characterizing the landscape of gene process dependencies in cancer

This study uses BioBombe to analyze gene process dependencies in cancer, identifying new therapeutic targets and drug candidates.

High-content live-cell time-lapse imaging predicts cells about to die via apoptosis

This paper uses high-content live-cell time-lapse imaging to predict apoptosis in HeLa cells using machine learning, detecting it 100 minutes after treatment with Staurosporine.

Stellar quality control for single-cell image-based profiling with coSMicQC

coSMicQC is a Python package for single-cell image-based profiling quality control, improving data quality by filtering low-quality cells and reducing technical noise.

High-content microscopy and machine learning characterize a cell morphology signature of NF1 genotype in Schwann cells

The paper uses high-content microscopy and machine learning to identify a cell morphology signature of NF1 genotype in Schwann cells, but model performance declines when applied to new cell lines.

A versatile information retrieval framework for evaluating profile strength and similarity

The paper introduces a statistical framework using mean average precision (mAP) to evaluate phenotypic activity and consistency in high-dimensional profiling data across various modalities and perturbation types.

A morphology and secretome map of pyroptosis

The paper characterizes pyroptosis using secretome profiling and morphology analysis, revealing complex connections with apoptosis through machine learning models.

A genome-wide atlas of human cell morphology

The paper presents a genome-wide perturbation atlas using CRISPR-Cas9 knockouts to link genes to cellular phenotypes, identifying TMEM251/LYSET as a Golgi-resident protein involved in lysosomal enzyme trafficking.

Combining NeuroPainting with transcriptomics reveals cell-type-specific morphological and molecular signatures of the 22q11.2 deletion

The paper introduces NeuroPainting, a morphological phenotyping method, and combines it with transcriptomics to identify cell-type-specific effects of the 22q11.2 deletion on mitochondrial structure and gene expression in astrocytes.

High-dimensional phenotyping to define the genetic basis of cellular morphology

This study identifies genetic variants associated with cellular morphology using high-content imaging and genomic data from iPSCs.

Discovering cellular programs of intrinsic and extrinsic drivers of metabolic traits using LipocyteProfiler

The paper introduces LipocyteProfiler, a high-throughput microscopy assay to profile lipid-storing cells and identifies cellular programs linked to metabolic traits like insulin resistance and lipodystrophy.

OpenPBTA: An Open Pediatric Brain Tumor Atlas

OpenPBTA provides a collaborative platform to genomically characterize 1,074 pediatric brain tumors and 22 cell lines, offering analysis modules and somatic datasets for research and clinical applications.

Morphology and gene expression profiling provide complementary information for mapping cell state

The paper compares morphology and gene expression data to determine their complementary roles in mapping cell states across drug perturbations.

Predicting cell health phenotypes using image-based morphology profiling

The paper develops a method to predict cell health phenotypes from Cell Painting images using machine learning, validated across multiple perturbations and assays.

Genome-wide association study implicates novel loci and reveals candidate effector genes for longitudinal pediatric bone accrual through variant-to-gene mapping

The paper identifies 40 novel genetic loci associated with longitudinal pediatric bone accrual and identifies candidate effector genes through variant-to-gene mapping and functional validation.

Epigenomic profiling of neuroblastoma cell lines

The paper provides ChIP-Seq and ATAC-Seq data for MYCN, MYC, and histone marks in neuroblastoma cell lines to study gene regulation and enhancer dynamics.

Genomic profiling of childhood tumor patient-derived xenograft models to enable rational clinical trial design

This paper genomically characterizes 261 pediatric cancer patient-derived xenograft models to guide rational clinical trial design and drug development.

Sequential compression of gene expression across dimensionalities and methods reveals no single best method or dimensionality

The paper evaluates multiple compression methods and dimensionalities on gene expression data, finding no single best approach but showing benefits from ensemble methods across different dimensions.

PathCORE-T: identifying and visualizing globally co-occurring pathways in large transcriptomic compendia

PathCORE-T identifies and visualizes globally co-occurring pathways across transcriptomic datasets by analyzing features from machine learning models.

Opportunities And Obstacles For Deep Learning In Biology And Medicine

Deep learning rivals or beats baselines across imaging, health records, genomics and drug discovery but is not yet transformative; scarce labels, privacy and opacity remain hurdles.

Functional network community detection can disaggregate and filter multiple underlying pathways in enrichment analyses

The paper develops a method combining functional network community detection with pathway enrichment to reduce false positives in complex gene sets.

Extracting a Biologically Relevant Latent Space from Cancer Transcriptomes with Variational Autoencoders

This paper trains a variational autoencoder on TCGA RNA-seq data to model cancer gene expression and identify biologically relevant latent features.

Deconvolution of DNA methylation identifies differentially methylated gene regions on 1p36 across breast cancer subtypes

Implicating candidate genes at GWAS signals by leveraging topologically associating domains

The paper introduces a method to identify candidate genes at GWAS signals by analyzing gene content within topologically associating domains, demonstrating its application to bone mineral density data.

A machine learning classifier trained on cancer transcriptomes detects NF1 inactivation signal in glioblastoma

The paper trains a machine learning classifier on cancer transcriptomes to detect NF1 inactivation in glioblastoma, showing it can predict NF1 status based on gene expression and mutation data.