Index of /wrpearson/fasta

[ICO]NameLast modifiedSizeDescription

[PARENTDIR]Parent Directory   -  
[TXT]COPYRIGHT 2014-11-25 15:45 1.1K 
[DIR]CURRENT/ 2023-06-03 10:41 -  
[DIR]OLD/ 2011-02-21 12:19 -  
[DIR]PUBS/ 2010-07-22 10:17 -  
[TXT]changes_v34.html 2023-06-03 10:40 14K 
[TXT]changes_v35.html 2023-06-03 10:40 8.6K 
[TXT]changes_v36.html 2023-06-03 10:40 23K 
[DIR]data/ 2016-02-20 14:41 -  
[DIR]executables/ 2023-06-03 10:39 -  
[DIR]fasta2/ 2008-04-09 05:55 -  
[   ]fasta3.tar.gz 2023-06-03 10:40 1.3M 
[DIR]fasta3/ 2023-06-03 10:36 -  
[DIR]fasta33-35/ 2017-05-31 08:10 -  
[DIR]fasta36/ 2023-06-03 10:36 -  
[   ]fasta_guide.pdf 2023-06-03 10:40 279K 
[TXT]fasta_versions.html 2023-06-03 10:40 4.8K 
[DIR]near_optimal/ 2005-01-20 14:55 -  
[DIR]other/ 2008-04-09 05:23 -  
[TXT]readme.COPYRIGHT 2014-11-25 15:46 1.1K 
[DIR]www/ 2014-04-29 12:39 -  

The FASTA package - protein and DNA sequence similarity searching and alignment programs

This directory contains the source code for the FASTA package of programs (W. R. Pearson and D. J. Lipman (1988), "Improved Tools for Biological Sequence Analysis", PNAS 85:2444-2448). The current verion of the program is fasta-36.3.8i.

If you are reading this at, links are available to executable binaries for Linux, MacOS, and Windows. The source code is also available from

The FASTA package offers many of the same programs as BLAST, but takes a different approach to statistical estimates, and provides additional optimal programs for local (ssearch36) and global (ggsearch36, glsearch36) alignment, and for non-overlapping internal local alignments (lalign36).

The programs available include:

FASTA BLAST description

fasta36 blastp/blastn Protein and DNA local similarity search
ssearch36 optimal Smith-Waterman search -- vectorized on Intel and Arm architectures
ggsearch36 optimal global Needleman-Wunsche search -- vectorized on Intel and Arm architectures
glsearch36 optimal global(query)/local (library) search -- vectorized on Intel and Arm architectures
fastx36 / fasty36 blastx DNA query search against protein sequence database. (fasty36 uses a slower, more sophisticated frame shift aligner)
tfastx36 / tfasty36 tblastn protein query search against DNA database

fastf36 / tfastf36 compares an ordered peptide mixture against a protein (fastf36) or DNA (tfastf36) database
fastm36 / tfastm36 compares a set of ordered peptide against a protein (fastf36) or DNA (tfastf36) database or oligonucleotides against a DNA database
fasts36 / tfasts36 compares an unordered set of peptides against a protein (fasts36) or DNA (tfasts36) database

lalign36 look for non-overlapping internal alignments, similar to a "dot-plot," but with statistical signficance

Changes in fasta-36.3.8i May, 2023

  1. restore the default -s BL62 gap penalties to -8, -1 (they were -11, -1, matching -s BP62)

  2. restore functionality of -A option, which forces Smith-Waterman final display alignments with DNA (normally banded Smith-Waterman is used) -XA forces banded alignments.

  3. add --id option to scripts/ to add a custom identifier

Changes in fasta-36.3.8i Nov, 2022

  1. bug fix to remove duplicate variant annotations

  2. update to scripts/ and annotation scripts.

  3. modify code to reduce mktemp compilation warning messages

  4. changes to annotation scripts for Pfam shutdown; new,

  5. a new option, r for -m 8CB that displays the raw optimal alignment score (typically Smith-Waterman).

Changes in fasta-36.3.8i Sept, 2021

  1. Enable translation table -t 9 for Echinoderms. This bug has existed since alternate translation tables were first made available.

Changes in fasta-36.3.8i May, 2021

  1. Add an option, -Xg, that preserves the gi|12345 string the score summary and alignment output.

Changes in fasta-36.3.8i Nov, 2020

  1. fasta-36.3.8i (November, 2020) incorporates the SIMDe (SIMD-everywhere, macro definitions that allow the smith_waterman_sse2.c, global_sse2.c, and glocal_sse2.c code to be compiled on non-Intel architectures (currently tested on ARM/NEON). Many thanks to Michael R. Crusoe ( for the SIMDE code converstion, and to Evan Nemerson for creating SIMDe.

  2. The code to read FASTA format sequence files now ignores lines with '#' at the beginning, for compatibility with PSI Extended FASTA Format (PEFF) files (

Changes in fasta-36.3.8h May, 2020

  1. fasta-36.3.8h (May 2020) fixes a bug that appeared when multiple query sequences were searched against a large library that would not fit in memory. In that case, the number of library sequences and residues increased by the library size with each new search.

  2. More consistent formats for ERROR and Warning messages.

  3. Corrections to code to address compiler warnings with gcc8/9.

  4. addition of 's' option to show similarity in -m8CBls (or -m8CBs, -m8CBsl) and 'd' option to show raw (unaligned) domain information.

Changes in fasta-36.3.8h February, 2020

  1. The license for Michael Farrar's Smith-Waterman sse2 code and global/glocal sse2 code is now open source (BSD), see COPYRIGHT.sse2 for details.

Changes in fasta-36.3.8h August, 2019

  1. Modifications to support makeblastdb format v5 databases. Currently, only simple database reads have been tested.

Changes in fasta-36.3.8h March, 2019

  1. Translation table 1 (-t 1) now translates 'TGA'->'U' (selenocysteine).

  2. New script for extracting DNA sequences from genomes (scripts/ Currently works with human (hg38), mouse (mm10), and rat (rn6).

Changes in fasta-36.3.8h January, 2019

  1. Bug fixes: fastx/tfastx searches done with the -t t option (which adds a * to protein sequences so that termination codons can be matched), did not work properly with the VT series of matrices, particularly VT10. This has been fixed.

  2. New features: Both query and library/subject sequences can be generated by specifying a program script, either by putting a ! at the start of the query/subject file name, or by specifying library type 9. Thus, fasta36 \\!../scripts/ /seqlib/swissprot.fa or fasta36 "../scripts/ 9" /seqlib/swissprot.fa will compare two query sequences, P09488 and P30711, to SwissProt, by downloading them from Uniprot using the script (which can download sequences using either Uniprot or RefSeq protein accessions). Often, the leading ! must be escaped from shell interpretation with \\!.

New scripts that return FASTA sequences using accessions or genome coordinates are available in scripts/.,, and can download either protein or mRNA RefSeq entries. retrieves a protein and its isoforms from a MySQL database. extracts genome sequences using coordinates from local reference genomes (hg38 and mm10 included by default).

Changes in fasta-36.3.8h December, 2018

The scripts/ and now include the option --gen_coord which provides the associated genome coordinate (including chromosome) as a feature, indicated by '<' (start of exon) and '>' (end of exon).

Changes in fasta-36.3.8h released November, 2018

fasta-36.3.8h provides new scripts and modifications to the fasta programs that normalize the process of merging sub-alignment scores and region information into both FASTA and BLAST results. To move BLASTP towards FASTA with respect to alignment annotation and sub-alignment scoring:

  1. The runs a blast search, finds and scores domain information for the alignments, and merges this information back into the blast output .html file. This script uses:

    1. --query query.file --ann_script --q_ann_script blast.btab_file > blast.btab_file_ann (a blast tabular file with one or two new fields, an annotation field and (optionally with --dom_info) a raw domain content field.
    2. --btab blast.btab_file_ann blast.html > blast_ann.html (merge the annotations and domain content information in the blast.btab_file_ann file together with the standard blast output file to produce annotated alignments.
    3. In addition, is available to rename exons (later other domains) in the subject sequences to match the exon labeling in the aligned query sequence.
    4. can be used to adjust color sets for homologous domains.
  2. There is also an equivalent script that provides similar funtionality for the FASTA programs. This script does not need to use to produce domain subalignment scores (that functionality is provided in FASTA), but it also can use and to modify the names of the aligned exons/domains in the subject sequences.

  3. To support the independence of the blastp/fasta output from html annotation, the FASTA package includes some new options:

    1. The -m 8CBL option includes query sequence length and subject sequence length in the blast tabular output. In addition, if domain annotations are available, the raw domain coordinates are provided in an additional field after the annotation/subalignment scoring field. -m 8CBl provides the sequence lengths, but does not add the raw domain coordinates.

    2. The -Xa option prevents annotation information from being included in the html output -- it is only available in the -m 8CB (or -m 8CBL/l) output

    3. To reduce problems with spaces in script arguements, annotation scripts with spaces separating arguments can use '+' instead of ' '.

    4. The script produces both a conventional alignment on stdout and a -m 8CBL alignment, which is sent to a separate file, which is separated from the -m F8CBL option with a =, thus -m F8CBL=tmp_output.blast_tab.

Changes in fasta-36.3.8g released 23-Oct-2018

  1. (Oct. 2018) Improvements to scripts in the psisearch2/ directory:

    1. psisearch2/

      1. the --clustal option produces a "CLUSTALW (1.8)", which is required for some downstream programs
      2. the --trunc_acc option removes the database and accession from identifiers of the form: sp|P09488|GSTM1_HUMAN to produce GSTM1_HUMAN.
      3. the --min_align option specifies the fraction of the query sequence that must be aligned (q_end-q_start+1)/q_length) Together, these changes make it possible for the output of to be used by the EMBOSS program fprotdist.
    2. A more general implementation of, which does psisearch2 one iteration at a time, and a new equivalent, which uses psiblast to do the search.

Changes in fasta-36.3.8g released 5-Aug-2018

  1. (Apr 2018) incorporation of -t t1 termination codes ("*") in -m 8CB, -m 8CC, and -m9C so that aligned termination codons are indicated as ** (-m8CB) or *1 (-m8CC, -m9C).

  2. (Mar 2018) Updates to scripts/ to provide subalignment scoring for blastp searches (BLOSUM62 only). (see doc/readme.v36)

  3. (Feb. 2018) a new extended option, -XB, which causes percent identity, percent similarity, and alignment length to be calculated using the BLAST model, which does not count gaps in the alignment length.

see readme.v36 for other bug fixes.

Changes in fasta-36.3.8g released 31-Dec-2017

  1. (December, 2017) -- Make statistical thresholds more robust for small E()-values with normally distributed scores (ggsearch36,glsearch36).

  2. (September, 2017) Treat lower-case queries with no upper-case residues as uppercase with -S option.

  3. (May, 2017) Improvements/fixes to sub-alignment scoring strategies.

  4. Improvements/fixes to psisearch2 scripts.

For more detailed information, see doc/readme.v36.