Fig. 2: The iCMod pipeline for analysis of multi-omics data and prediction of circadian protein kinases.

We constructed a reference database for MS/MS searching containing only protein sequences with corresponding mRNA levels of FPKM ≥ 1 in at least one sample. After the database search for raw proteomic and phosphoproteomic spectra, the intensity values of proteins were normalized to one for each sample. Phosphorylation level was normalized to the corresponding protein level. ARSER was used to identify cycling mRNAs, proteins, and NCPs. GPS 2.1 was subsequently employed to predict ssKSRs, followed by a two-sided hypergeometric test to predict potential circadian kinases.