I notice that tg_index has some "hidden" support for the AFG file format. I want to explore this somewhat and have some comments:
1) Contrary to what is in the tg_index.c file, the AFG file format is not a Velvet format but is the AMOS message format:
2) AMOS provide a c++ API to this format which could be used by gap5 for AFG I/O:
AMOS has some good tools for piecing together data about an assembly e.g. libraries, scaffolds, read pairing info etc and generating an AFG message file. AMOS uses this message file as an interchange format and uses it to generate an AMOS bank (using bank-transact) on which most AMOS tools subsequently work. Individual messages can be used to update entries in the AMOS bank.
If tg_index could more fully support the AFG format, people could easily convert a full assembly in AFG format into a gap5 db. This would allow people to use the visualisation aspects of AMOS hawkeye to identify potential missassemblies and use gap5 to make edits. Ideally, gap5 would be able to take individual massages in AFG format to update its own internal data structures and output chages in AFG format. This would allow people to streamline the visualisation/editing of an assemby using the AFG format as a messaging system between AMOS bank and Gap5 db. I think the only thing that would be needed would be to map AFG to gap5's internal data structures and use the AMOS c++ API to read/write AFG format messages.
Log in to post a comment.