Non-EST based prediction of exon skipping and intron retention events using Pfam information

Nucleic Acids Res. 2005 Oct 4;33(17):5611-21. doi: 10.1093/nar/gki870. Print 2005.


Most of the known alternative splice events have been detected by the comparison of expressed sequence tags (ESTs) and cDNAs. However, not all splice events are represented in EST databases since ESTs have several biases. Therefore, non-EST based approaches are needed to extend our view of a transcriptome. Here, we describe a novel method for the ab initio prediction of alternative splice events that is solely based on the annotation of Pfam domains. Furthermore, we applied this approach in a genome-wide manner to all human RefSeq transcripts and predicted a total of 321 exon skipping and intron retention events. We show that this method is very reliable as 78% (250 of 321) of our predictions are confirmed by ESTs or cDNAs. Subsequent analyses of splice events within Pfam domains revealed a significant preference of alternative exon junctions to be located at the protein surface and to avoid secondary structure elements. Thus, splice events within Pfams are probable to alter the structure and function of a domain which makes them highly interesting for detailed biological investigation. As Pfam domains are annotated in many other species, our strategy to predict exon skipping and intron retention events might be important for species with a lower number of ESTs.

Publication types

  • Evaluation Study
  • Research Support, Non-U.S. Gov't

MeSH terms

  • Algorithms
  • Alternative Splicing*
  • Amino Acid Sequence
  • Computational Biology / methods
  • Exons*
  • Expressed Sequence Tags
  • Genomics / methods*
  • Humans
  • Introns*
  • Molecular Sequence Data
  • Peptides / genetics
  • Protein Structure, Tertiary


  • Peptides