UniPath

UniPath computes pathway and gene-set enrichment scores from single-cell open‑chromatin (ATAC-seq) and transcriptome (RNA-seq) read-counts to quantify cellular heterogeneity and support pathway-level analyses.

Key Features:

• Pathway and Gene-Set Enrichment Scores: Represents each cell by pathway and gene-set enrichment scores computed from transformed open‑chromatin or gene‑expression profiles using a robust statistical method to ensure accuracy, consistency, and scalability.
• Handling Variability and Drop-Out Rates: Implements statistical handling of variability and drop-out rates in read-counts to reduce artifacts arising from systematic patterns.
• Dimension Reduction: Provides an alternative approach for dimension reduction of single-cell open‑chromatin profiles.
• Pseudo-Temporal Ordering: Predicts pseudo-temporal ordering of cells using pathway enrichment scores and can suppress covariates to improve chronological arrangement.
• Comparative Analysis of Cell Populations: Compares two groups of cells by exploiting pathway co-occurrence to infer context-specific regulations.

Scientific Applications:

• Single-Cell RNA-seq and ATAC-seq: Applicable to single-cell RNA-seq (RNA-seq) and ATAC-seq profiles and scalable to atlas-scale datasets.
• Biologically Meaningful Co-Clustering: Facilitates discovery of biologically meaningful co-clustering across distant organs, as demonstrated on the mouse cell atlas.
• Inference of Context-Specific Regulations: Enables inference of context-specific regulatory programs, including applications in cancer research.

Methodology:

Transforms open‑chromatin or gene‑expression read-counts into pathway/gene-set enrichment scores using robust statistical methods; handles variability and drop-out rates; applies an alternative dimension reduction for open‑chromatin; predicts pseudo-temporal ordering from pathway scores with covariate suppression; and compares cell groups via pathway co‑occurrence.