Differential Expression Analysis

Differential expression analysis is performed for miRNA families, precursor groups, miRNAs, and piRNAs. If novel precursor discovery is enabled, the precursor groups include both known and novel precursor groups.

MiRNAs include mature miRNAs and isomiRs (know precursor). If novel miRNA prediction is enabled, miRNAs also include novel mature miRNAs and isomiRs (novel precursor).

DESeq2 filters out low expressed markers with normalized mean count < 10 before testing³. The DESeq2 variance model is used to detect and exclude outliers based on extreme variation between replicates. The status (filtered or passed) and the result of the analysis (mean expression, fold change, standard error, p-value, q-value, etc.) is reported for each marker.

If novel precursor discovery is enabled, the novel precursors predicted for the 2 sample groups in each pairwise differential expression analysis are merged. The merged novel precursor ID format is "{control group}_vs_{comparison group}-novelMiR-{nonnegative integer}:{ID1}[+ID2[+ID3...]]", where ID1, ID2 and ID3 are the IDs of the merged member novel precursors. Because the first part of the ID uniquely identifies a merged novel precursor, the member novel precursor IDs are omitted in reports.

Similarly, novel precursors are merged across all the sample groups to generate global plots, such as the expression heat map. The merged novel precursor ID format is "global-novelMiR-{nonnegative integer}:{ID1}[+ID2[+ID3...]]". The member precursor IDs are also omitted in reports.