FitDevo

FitDevo infers single-cell developmental potential from scRNA-seq data by generating sample-specific gene weights (SSGW) to evaluate developmental stages and molecular signatures.

Key Features:

• Sample-specific gene weights (SSGW): Generates SSGW for each sample to quantify gene contributions to developmental potential.
• Generalized linear model (GLM): Produces SSGW using a GLM that integrates sample-specific information and gene weights derived from a training set of 17 published scRNA-seq datasets.
• Correlation-based inference: Infers developmental potential by calculating the correlation between SSGW and individual gene expression profiles.
• Benchmark validation: Validated on a testing set of scRNA-seq data from 28 published datasets and reported superior performance compared with existing methods.
• Downstream analysis applicability: Applied to deconvolution analysis and spatial transcriptomic contexts to link developmental potential with tissue and disease states.

Scientific Applications:

• Epidermis deconvolution: Deconvolution analysis of epidermis tissue to resolve cellular contributions related to developmental potential.
• Spatial transcriptomics: Analysis of spatial transcriptomic data from hearts and intestines to map developmental potential across tissue architecture.
• Breast cancer studies: Investigation of developmental potential within breast cancer datasets to characterize tumor cell states.

Methodology:

FitDevo generates sample-specific gene weights through a generalized linear model that integrates sample-specific information and gene weights from a training dataset of 17 scRNA-seq datasets, then infers developmental potential by computing the correlation between SSGW and individual gene expression profiles, with performance evaluated on a testing set of 28 scRNA-seq datasets.