So far, the annotation of translation initiation sites (TISs) has been based mostly upon bioinformatics rather than experimental evidence. We adapted ribosomal footprinting to puromycin-treated cells to generate a transcriptome-wide map of TISs in a human monocytic cell line. A neural network was trained on the ribosomal footprints observed at previously annotated AUG translation initiation codons (TICs), and used for the ab initio prediction of TISs in 5062 transcripts with sufficient sequence coverage. Functional interpretation suggested 2994 novel upstream open reading frames (uORFs) in the 5′ UTR, 1406 uORFs overlapping with the coding sequence, and 546 N-terminal protein extensions. The TIS detection method was validated on the basis of previously published alternative TISs and uORFs. Among primates, TICs in newly annotated TISs were significantly more conserved than control codons, both for AUGs and near-cognate codons. The transcriptome-wide map of novel candidate TISs derived as part of the study will shed further light on the way in which human proteome diversity is influenced by alternative translation initiation and regulation.

Thousands of tandem alternative splice sites (TASS) give rise to mRNA insertion/deletion variants with small size differences. Recent work has concentrated on the question of biological relevance in general, and the physiological regulation of TASS in particular. We have quantitatively studied 11 representative TASS cases in comparison to one mutually exclusive exon case and two cassette exons (CEs) using a panel of human and mouse tissues, as well as cultured cell lines. Tissues show small but significant differences in TASS isoform ratios, with a variance 4- to 20-fold lower than seen for CEs. Remarkably, in cultured cells, all studied alternative splicing (AS) cases showed a cell-density-dependent shift of isoform ratios with similar time series profiles. A respective genome-wide co-regulation of TASS splicing was shown by next-generation mRNA sequencing data. Moreover, data from human and mouse organs indicate that this co-regulation of TASS occurs in vivo, with brain showing the strongest difference to other organs. Together, the results indicate a physiological AS regulation mechanism that functions almost independently from the splice site context and sequence.