Figure 2.

Many human alternative isoforms in SWISS-PROT derive from PTC⁺ mRNAs. (a) We analyzed each of the human SWISS-PROT entries containing a VARSPLIC line in its feature table, using this information to assemble protein isoform sequences. Ambiguous VARSPLIC entries led us to discard five entries from our analysis at this point. (b) We next identified cDNA/mRNA sequences corresponding to each protein isoform assembled from SWISS-PROT. BLAST was used to align each protein isoform sequence to translated cDNA/mRNA sequences in GenBank and Refseq, filtering to ensure only high confidence matches. To obtain the coding sequence of each mRNA/cDNA sequence, we used LocusLink to map each to the correct human genomic contig sequence from the NCBI human genome build 30. We referred to the CDS feature of each GenBank or RefSeq cDNA/mRNA record to identify stop codon locations. (c) We used the SPIDEY mRNA-to-genomic DNA alignment program to determine the gene structure of each mRNA/cDNA isoform sequence. After generating these gene structures, we could determine the PTC⁺ status on the basis of stop codon location relative to exon-exon junctions. If the termination codon was found to be more than 50 nucleotides upstream of the final intron, the transcript was deemed PTC⁺ and designated a candidate target of NMD according to the model of mammalian PTC recognition. (d) Each putative PTC⁺ isoform was manually inspected for errors in gene structure prediction. These errors include false exon predictions due to poly(A) tails and cDNA/mRNA sequence not seen in the corresponding genomic sequence.