Internal organisation of CTS ASE

The cell type specific (CTS) allele specific expression (ASE) module is separated into three sub-modules:

1. ASreads
2. ASEperSNP
3. ASEperRegion

ASreads

This sub-module is able to quantify the number sequencing reads that overlap an allele at heterozygous SNP sites, by integrating any quantitative sequencing experiment (like RNA-seq or proseq) with genotype information.

The data created in this module can be used in the later modules for the quantification of allele specific expression.

ASEperSNP

This sub-module combines multiple ASfiles (Produced previously in the ASreads sub-module) of samples, and determines the probability of allelic imbalance per SNP. This module uses multiple methods of finding the allelic imbalance using a likelihood ratio test (LRT):

• Binomial LRT
• Beta binomial LRT

When dealing with heterogeneous tissue, and cell proportions are available:

• CTS Binomial LRT
• CTS beta binomial LRT

Per SNP, a P value is determined. The P value is the probability of finding no ASE. Lower P values mean more evidence for ASE at this SNP. Usually the Beta Binomial (BB) test is better at correction for sample specific biases, thus, the results of the Beta Binomial should usually be considered to be more accurate.

ASEperRegion

This sub-module combines multiple ASfiles (Produced previously in the ASreads submodule) of samples, and determines the probability of allelic imbalance per genomic region. This module uses multiple methods of finding the allelic imbalance using a likelihood ratio test (LRT):

• Binomial LRT
• Beta binomial LRT

Per Region, and test SNP, a P-value is determined, in the same was as the ASEperSNP module.