Abstract:
|
Differential expression analysis of RNA-seq can often be more challenged due to the presence of many covariates associated with experimental design, experimental units, RNA samples, etc. Ignoring relevant covariates or including irrelevant covariates can decrease the ability to detect differentially expressed genes. In this talk, we investigate the performance of backward selection and forward selection in an experiment that involves the measurement of transcript levels in blood samples taken from two lines of pigs that are divergently selected for residual feed intake. A simulation study shows advantages of backward selection and forward selection methods over the alternative approaches that either including or excluding all covariates.
|