FASTA program discussion list

[William Pearson <>]

A new version of the FASTA package, fasts-36.3.6f, is available from:

http://faculty.virginia.edu/wrpearson/fasta/fasta36/fasta-36.3.6f.tar.gz

This version has a number of improvements, and some bug fixes.

Improvements:

(1) The -XI option causes 100% identity to be shown only when there are no 
mismatches.  Previously, a sequence that had one mismatch in a 10,000 residue 
alignment would be shown as 100% identical because of rounding from 99.99% -> 
100.0%. With -XI, it would be shown as 99.9% identical.

(2) The -m 9c and -m9C options, which provide an encoding of the alignment 
(-m 9C provides a CIGAR string) have been extended to include -m 9d and -m9D, 
which add mismatches to the CIGAR encoding.  The -m 9c/C encoding showed 
matches (aligned residues, identical or not), insertions, and deletions.  The 
-m 9d/D encoding adds ‘X’ for mismatch to ‘M’ for match (CIGAR string).  This 
is a useful option for highly identical DNA sequences, but it can produce 
very long encodings for low identity protein alignments.

(3) lalign36 now allows the "-m 9/-m 9i/-m 9c/d/C/D" options, which display a 
summary of the high-scoring non-identical alignments, similar to the “The 
best scores are:” summary for database searches. For lalign36, the summary 
header is either “The best non-identical alignments are:” (default), or “The 
best alignments are:” (if the -XJ option is used to include the identity 
alignment).  The output format is unchanged unless a “-m 9” option is 
specified.

(4) Annotation scripts are provided to collect domain and site information 
from Uniprot (scripts/ann_feats_up_www2.pl) or domain information from Pfam 
(scripts/ann_pfam_www.pl) using the UniProt/EBI gff2 server or the Pfam web 
server.

(5) Code has been added to read the ambiguity array for NCBI 
blastdbfmt/formatdb formatted DNA libraries.

(6) The bit score calculation has been modified for sub-alignment scoring so 
that sub-alignment bit-scores add up to the total alignment bit-score.

(7) The PSSM ASN.1 parsing code has been updated to correctly parse psiblast+ 
ASN.1

As always, please let me know if you encounter problems.

Bill Pearson