ExCluster

ExCluster robustly detects differentially expressed exons between two conditions of RNA-seq data by flattening Ensembl and GENCODE GTF files into GFF files, counting reads per non-overlapping exon bin from BAM files, normalizing library sizes across biological replicates, and comparing two conditions to detect significantly differentially spliced genes with false discovery rates.

Key Features:

• GTF flattening: ExCluster flattens Ensembl and GENCODE GTF files into GFF files to define non-overlapping exon bins.
• Read counting: Read counting per non-overlapping exon bin from BAM files is performed using the featureCounts function from the Rsubread package.
• Library size normalization: Library sizes are normalized across all biological replicates prior to differential analysis.
• Two-condition comparison: ExCluster accepts exactly two conditions at a time and requires at least two independent biological replicates per condition.
• Significance testing: ExCluster produces false discovery rates (FDRs) per gene to detect significantly differentially spliced genes.
• Visualization: Exon log2 fold change (log2FC) means and variances may be plotted for each significantly differentially spliced gene.

Scientific Applications:

• Differential splicing detection: ExCluster identifies significantly differentially spliced genes between two conditions of RNA-seq data.
• RT-qPCR validation targeting: Exon log2FC plots help scientists develop hypotheses and target differential splicing events for RT-qPCR validation in the wet lab.

Methodology:

ExCluster flattens Ensembl and GENCODE GTF files into GFF files, counts reads per non-overlapping exon bin from BAM files using featureCounts from Rsubread, normalizes library sizes across all biological replicates, and compares two conditions to detect significantly differentially spliced genes with false discovery rates.