Publication at Faculty of Science, Central Library of Charles University |
2020
Abstract
The merit of RNASeq data relies heavily on correct normalization. However, most methods assume that the majority of transcripts show no differential expression between conditions.
This assumption may not always be correct, especially when one condition results in overexpression. We present a new method (NormQ) to normalize the RNASeq library size, using the relative proportion observed from RT-qPCR of selected marker genes.
The method was compared against the popular median-of-ratios method, using simulated and real-datasets. NormQproduced more matches to differentially expressed genes in the simulated dataset and more distribution profile matches for both simulated and real datasets.