version 0.22: important improvements to genomewide binaries (maintenance release)

[COTRASIF]

Changes in genomewide binaries since version 0.21:

1. FASTA files with line lengths different from 60 are now processed correctly. Previously, FASTA files with line lengths other than 60 had errors in absolute coordinates reported.
2. Non-existing input files (seqfile, infile, pfmfile) were causing a segfault.
3. It is now possible to specify --cutoff=0 instead of a workaround --cutoff=0.0001.
   
4. If PFM file had non-digit and/or non-space symbols, PFM parser misbehaved badly (for example, on PFMs copy-pasted from JASPAR - including square brackets). Now PFM parser performs input cleanup, and correctly processes PFMs with extra symbols.
5. If a TFBS with a score higher than cutoff was found exactly at the end of the sequence line in the FASTA file, it was producing duplicate result lines.
   

Downloading newer version is highly recommended.

Other fixes:

• orthology filter used to skip valid results in rare marginal cases
• more automated tests were developed to check for proper support for non-Ensembl FASTA files
• other minor fixes to internals

I would like to thank Iwe Muiser for help in identifying the problems with genomewide binaries, which were fixed in this release.

[Bogdan]

> Changes in genomewide binaries since version 0.21:

A small note on the "beta" status of genomewide binaries.

Given the scope of current automatic tests, "beta" tag will probably
be removed from genomewide binaries in version 0.24 (assuming that
complementary strand scanning will be added in version 0.23).