EMBOSS: Project Meeting (May 8th 2000) |
There is a related problem with sequence output, where "-ossingle" still creates the original output file even if it is not going to be used. "-ossingle" should leave the sequence code to make up the output file names.
Alan has added fuzztran to search for protein patterns in translated DNA sequences.
Norway have added lindna and linplasmid for linear DNA displays.
Ian has added the Norwegian changes to findkm to use a Hanes Woolf plot rather than Lineweaver Burk.
Peter has modified dbigcg to index PIR format database. They failed before because dbigcg looked for the extra length and date fields in GCG format.
dbiblast had a problem in reading some larger databases. This has been traced to a non-static array definition and is fixed.
Peter is working through the final details of the code to read dbiblast indexed databases.
Alan will make a new function, based on ajFileNewInlist, for processing wildcard files in a directory.
Two versions of ajSeqMakeUsa are available. One for generating USAs to be included in output, and another for internal USAs with file offsets for speed. There is much further testing required, and a few known cases where the generation of USAs remains problematic.
Richard will make another attempt at a Windows NT build.
For Release 1.0 the main outstanding tasks are Gary's documentation review and Peter's final changes to sequence reading. After release 1.0 we should be careful about the timing of releases to the FTP server so that version numbers and major changes can be controlled. Developers can get the latest source code from the CVS servers.
Plans for Release 2.0 can be discussed next week.
Peter has reviewed the log files and counted 600 users at HGMP, using over 90 different applications.
Feature output can be implemented in release 2.0 as the default output for most applications. Before then it can be added as an optional extra to applications.
Reading feature tables from sequence database entries has still to be implemented. Peter will add the code while reviewing sequence reading.
Peter will look into revising the format of the Web pages to be much plainer for ease of generating a tar distribution file, and to simplify porting the pages to the HGMP web server.
Peter will create an interfaces web page to document the interfaces that are being actively worked on and the progress to date.
Rodger Staden's group are making good progress on integration of graphics and ACD.
Peter is giving an HGMP seminar on EMBOSS tomorrow.
Next meeting on Monday 15th May, 11:00am, usual place.