Identification of Small Open Reading Frame-encoded Proteins in the Human Genome

One of the main goals of the Human Genome Project is to identify all protein-coding genes. There are ∼ 20,500 protein-coding genes annotated in the human reference databases. However, in the last few years, proteogenomics studies have predicted thousands of novel protein-coding regions, including low-molecular-weight proteins encoded by small open reading frames (sORFs) in untranslated regions of messenger RNAs and non-coding RNAs. Most of these predictions are based on bioinformatics analyses and ribosome footprint data. The validity of some of these sORF-encoded proteins (SEPs) has been established through functional characterization. With the growing number of predicted novel proteins, a strategy to identify reliable candidates that warrant further studies is needed. In this study, we developed an integrated proteogenomics workflow to identify a reliable set of novel protein-coding regions in the human genome based on their recurrent observations across multiple samples. Publicly available ribosome profiling and global proteomic datasets were used to establish protein-coding evidence. We predicted protein translation from 4008 sORFs based on recurrent ribosome occupancy signals across samples. In addition, we identified 825 SEPs based on proteomic data. Some of the novel protein-coding regions identified were located in genome-wide association study (GWAS) loci associated with various traits and disease phenotypes. Peptides from SEPs are also presented by major histocompatibility complex class I (MHC-I), similar to canonical proteins. Novel protein-coding regions reported in this study expand the current catalog of protein-coding genes and warrant experimental studies to elucidate their cellular functions and potential roles in human diseases.