Previous versions of source code

Note: There is a slight change in the GMAP database format starting with the GSNAP source code in 2010. If you wish to run older versions of GMAP on the newer databases, you will need to make a symbolic link from *.ref3offsets to *.idxoffsets, and from *.ref3positions to *.idxpositions.

• Current version 2007-09-28 (http://www.gene.com/share/gmap/src/gmap-2007-09-28.tar.gz)
  Major changes:
  • Implemented iit_update and gmap_update to update IIT files to version 2 (introduced in the 2007-07-16 version; see notes in the previous versions page for details), or convert them back to version 1.
  • GMAP can now read from multiple sequence files on the command line, like this: gmap -d gmapdb *.fa
  • Some fixes for special alignment cases, mostly affecting low-identity or cross-species alignment
  • Greatly reduced amounts of pre-allocated memory, which affected computers with low amounts of RAM
  • Get-genome can now access from map files entries with a particular direction or tag
  • Added a pruning step in the mapping stage
  • Fix for a rare divide-by-zero error
  • Fixed reading of input sequences with multiple PC line feeds
  • Turned off sorting of chromosomes in gmap_setup -S mode, which treats each contig as a chromosome
• Current version 2007-06-04 (http://www.gene.com/share/gmap/src/gmap-2007-06-04.tar.gz)
  Major changes:
  • Allowing skipping of internal exons that don't match genome (e.g., in cross-species alignments)
  • Changed meaning of -X flag. Now indicates heavy favoring of canonical and semi-canonical introns
  • Added coverage and identity information to GFF3 output
  • Fixed GFF3 output so protein sequence is in-frame
  • Fixed PSL output so query and target gaps are consistent with block starts and lengths
  • Better parsing of NCBI genome headers in gmap_setup
  • Improvements in approximate alignment algorithm
  • Various bug fixes, including some fatal bugs and excessive memory usage in some cases
• version 2006-12-18 (http://www.gene.com/share/gmap/src/gmap-2006-12-18.tar.gz)
  Major changes:
  • Improvements in aligning low-identity sequences
  • Pruning of spurious exons at ends of alignment, and added -k flag to gmap to control this
  • Fixed minor bugs in translations by gmap
  • Now able to handle nucleotide U (representing T) in query sequence
  • Made strict translation of cDNA sequence the default. Added -Y flag to gmap to perform tolerant translation that corrects for frameshifts
  • Added -p flag to gmap to control pruning of poor or repetitive sequences
  • Change to output format of Coverage field, which now reports query length and trimmed query coordinates
  • Support for GFF files: GMAP can now output in GFF gene format (-f 2 flag) and match format (-f 3 flag), and the IIT utilities can now parse GFF files
  • New output type for gmap: IIT FASTA format (-f 8 flag)
  • GMAP mapping facility (-m flag) can now use IIT files with chromosomal coordinates and tags, in addition to those with universal coordinates and strand tags
  • Added printing of chromosomal subsets by get-genome
  • Added -u flag to get-genome and iit_get to print flanking intervals
  • Provided long options and documentation, including --help, for IIT utilities
  • Fixed bug in iit_get for 32-bit machines which truncated coordinates at 31 bits
  • Improved parsing of .md files in gmap_setup
• version 2006-04-21 (http://www.gene.com/share/gmap/src/gmap-2006-04-21.tar.gz)
  Major changes:
  • Improvements in speed, sensitivity, and alignment quality
  • Added intron lengths to exon summary output
  • Changes in output format for reporting mutations (now called "Amino acid changes").
  • Added printing of chromosomal subset
  • Provided greater user control in limiting intron lengths
  • Simplified setup process for creating new GMAP databases. Now gmap_setup creates a Makefile.
  • Fixed problems with PSL output coordinates
  • Changed output in compressed output from "dir:unknown" to "dir:indet".
  • Improved map-only mode (-1 flag)
• version 2005-10-25 (http://www.gene.com/share/gmap/src/gmap-2005-10-25.tar.gz)
  Major changes:
  • Created two types of batch mode: one for indices only, and one for indices and genome. The -B flag now takes an argument.
  • Added instructions and options for gmap_setup to create a full, uncompressed genome, containing non-standard or lower-case characters
  • Fixed psl (BLAT) output format to create new blocks at indels
  • Improved handling of non-standard and lower-case characters, with ambiguous matches now being marked in the alignment
  • Added large file support during compilation
  • Added an option in fa_coords to treat each sequence as a separate chromosome
  • Changed gmap_setup to generate a Makefile
  • Fixed bug in handling chromosomes greater than 2 gigabytes in length
  • The intron gap parameter (-i flag) can now be arbitrarily large, and shows the entire intron if necessary
  • Added option to look up map IIT information for each exon
  • Get-genome can now look up map IIT files
  • Restored printing of strain information, if available
  • Added sense/antisense information to compressed output
• version 2005-07-15 (http://www.gene.com/share/gmap/src/gmap-2005-07-15.tar.gz)
  Major changes:
  • Added psl (BLAT) output format
  • Added a range format for get-genome
  • Fixed bug in cross-species alignment
  • Fixed bug in iterative handling of short exons
• version 2005-06-23 (http://www.gene.com/share/gmap/src/gmap-2005-06-23.tar.gz)
  Major changes:
  • Improvements to cross-species alignments
  • Introduced option for alignment to be truncated at coding region
  • Turned off default identification of microexons at sequence ends
• version 2005-05-20 (http://www.gene.com/share/gmap/src/gmap-2005-05-20.tar.gz)
  Major changes:
  • Fixed bug when multiple threads read a file that was not memory mapped
  • Database setup now properly handles contigs marked as reverse complement
  • Speed improvements to nucleotide-level dynamic programming
• version 2005-05-06 (http://www.gene.com/share/gmap/src/gmap-2005-05-06.tar.gz)
  Major changes:
  • Added check for repetitive cDNA sequences
  • Using oligomer-based method for trimming sequence ends
  • Indexing intervals other than 6 can be specified during database setup
• version 2005-03-11 (http://www.gene.com/share/gmap/src/gmap-2005-03-11.tar.gz)
  Major changes:
  • Introduced concept of chromosomal subsets
  • By default, get-genome now prints just reference strain
  • List of contigs now indicates if reverse complement was needed
  • Mapping algorithm uses dangling matches to determine whether to sample further