FrameshiftFinder

Work in progress!

• concept to input nucleotide sequence file containing at least two sequences in fasta format.
• Translate nucleotide sesquences to protein sequence (using correct translation table).
• Align both nucleotide and protein sequences.
• Score alignments
• Iterate through the nucleotide alignment looking for gaps < 3 bp long (i.e. possible frameshift)
• At each gap site, delete one nucleotide, retranslate and realign the protein translation.
• Score the new protein alignment and compare it to the orginal. If the revised alignment is better report this as a potentional frameshift.
• Alternatively if the alignment is worse delete two nucleotides at the orginal gap site and try gain.
• If the revised alignment is better report this as a potentional frameshift.
• If not move on to the next gap.
• Continue for all sequences in alignment.
• Need to consider internals stops (maybe ignored in the alignment?)
• Need to consider how this scales up to large inputs. May want to remove identical sequences or cluster the sequences to reduce the input file size.

example files

• input file: missingcytb.fsa
• translated protein sequences: missingcytb.pro
• alignments in clustal format: misssingcytb_nuc.aln and missingcytb_pro.aln