Identification and analysis of ancestral hominoid transcriptome inferred from cross-species transcript and processed...

Comparative transcriptomics studies in hominoids are difficult because of lack of EST information in the great apes. Nevertheless, processed pseudogenes (PPGs), which are reverse-transcribed ancient transcripts present in the current genome, can be regarded as a virtual transcript resource that may compensate for the paucity of ESTs in non-human hominoids. Here we show that chimpanzee PPGs can be applied to identification of novel human exons/alternatively spliced variants (ASVs) and inference of the ancestral hominoid transcriptome and chimpanzee exon loss events. We develop a method for comparatively extracting novel transcripts from PPGs (designated "CENTP") and identify 643 novel human exons/ASVs. RT-PCR-sequencing experiments confirmed >50% of the tested exons/ASVs, supporting the effectiveness of the CENTP pipeline. With reference to the ancestral transcriptome inferred by CENTP, 47 chimpanzee exon loss events are identified. Furthermore, by combining out-group and PPG information, we identify 20 chimpanzee-specific exon loss and 10 human-specific exon gain events. We also demonstrate that the ancestral transcriptome and exon loss/gain events inferred based on comparisons of current transcripts may be incomplete (or occasionally inappropriate) because ancestral transcripts may not be represented in the ESTs of existing species. Finally, functional analysis reveals that the novel exons identified based on chimpanzee transcripts are significantly enriched in genes related to translation regulatory activity and viral life cycle, suggesting different expression levels of the associated transcripts, and thus divergent splicing isoform composition between human and chimpanzee in these functional categories.