WemIQ

WemIQ quantifies isoform and gene expression levels and exon splicing (exon inclusion) ratios from bulk and single-cell RNA-seq data using a weighted-log-likelihood expectation-maximization approach and generalized Poisson bias correction.

Key Features:

• Weighted-Log-Likelihood Expectation Maximization (EM): Employs a weighted-log-likelihood EM algorithm to distribute sequencing reads among isoforms for precise isoform and gene expression quantification and exon inclusion rate estimation.
• Bias Correction via Generalized Poisson Model: Estimates weights representing oversampling or undersampling of reads using a generalized Poisson model without assuming specific sources or formats of bias.
• Robust Expression Estimates: Produces consistent isoform, gene expression, and exon inclusion rate estimates across different laboratories and protocols.
• Single-Cell RNA-seq Application: Distinguishes bias heterogeneity from true biological variability to reveal smaller cell-to-cell expression differences in single-cell RNA-seq data.

Scientific Applications:

• Isoform Expression Analysis: Quantifies isoform levels to facilitate studies of alternative splicing and isoform-level gene regulation.
• Exon Splicing Studies: Derives exon inclusion rates to investigate splicing patterns and their contribution to protein diversity.
• Single-Cell Transcriptomics: Separates technical bias from biological heterogeneity to support analysis of cellular phenotypes and states in single-cell RNA-seq datasets.

Methodology:

Uses a weighted-log-likelihood expectation-maximization algorithm to assign reads to isoforms and estimate isoform/gene expression and exon inclusion rates, and estimates sampling bias weights via a generalized Poisson model without assuming bias sources or formats; applies the same approach to single-cell RNA-seq to separate bias heterogeneity from biological variability.