DEAr – Differential Expression Analyzer

Abstract

Abstract Genes expressions are key features of cells and tissues studied in laboratories and clinics. Before analyzing expression data, normalization must be performed. Normalization methods can be generalized to the idea of forming a baseline from a subset of stable genes and then representing gene expressions as differences from this baseline. There are solutions available to help verify that the genes provided are stable enough to be used as a reference. Such a strategy works for high-throughput sequencing, but in cases of real-time PCR we have to work with a limited number of genes and especially when we compare with data from previous experiments. In such cases, we may struggle to find genes that are stable for the required cells and conditions. Here I present a program that simultaneously checks expressions of all genes in the data set and selects the most stable as a baseline for the less stable. The Differential Expression Analyzer (DEAr) assigns weights to all values depending on accuracy of direct measurements and reproducibility of differential expressions in the dataset, at the same time the program utilizes external knowledge about stability or instability of genes in considered materials and contributes it to weights as well. The important improvement is that the original algorithm allows individual weight to be assigned to each gene expression value for each sample to perform normalization. It allows DEAr to work without input for possible missing values. The algorithm is based on recursive computations that are described in detail in the article. DEAr is packaged in an executable file for the OS Windows. DEAr accepts and returns data in Excel format, so no programming skills are required for fast, automated analysis of differential expression.