EMBOSS: Project Meeting (Mon 30th March 09) |
Peter has updated pepwindow and pepwindowall to correctly process the sequence start and end positions. Both applications were missing the final residue. This is now fixed. The amino acid data files can be (re)normalized to have a mean of zero and a standard deviation of 1.0 to give consistent results. Some of the Nakai values (from aaindexextract) are unnormalized. Note that renormalizing a normalized set of values can give slightly different numbers. The graph title of both programs is made more general - it assumed hydrophobicity values.
Peter has modified applications that process patterns (prosite or regular expression) so that the suffix "1" is no longer added where there is only one pattern.
Mahmut reported emma can fail to write a DND file if a SoapLab job has many sequences with the same name.
Mahmut has looked into Ensembl DAS stylesheets.
Peter has fixed the alignment code to use double precision calculation. For long alignments it was possible to lose track of the path at single precision. Function results are also changed to return double precision score values.
The recent hit count is 2m/month, though over 75% are in the load balancer. About 500k are real hits. Many of these are Taverna plugin startups requesting service metadata. The LSF job count shows many hits are real jobs running.
Alan has built a new mEMBOSS with the alignment algorithm patch. The bundlewin utility failed to find exported functions if there is a tab at the start of the prototype in the header file.
Peter has updated the "command line data qualifiers" section. Examples were changed from EMBL to SwissProt so that ID and accession number are different (EMBL changed policy on identifiers since the original documentation was written).
Editors for the book text were discussed. Emacs was preferred to XMLmind.
Alan has modified eprimer3 to report the original output. The application is parsing and rewriting the output. An option to produce the parsed results was added.
The meaning of "reversed" in diffseq output needs checking. An insertion in the second sequence can give strange results.
Mahmut also reported that dotmatcher with a short second sequence crashes with an unsigned integer condition. This will need a general check through all applications for similar problems.