Merge two gff3 files?

[BobFreemanMA]

I have two gff3 files, one of primary transcripts and one of alternative transcripts, that need to be merged into one file. Since the files must be interwoven (some gene IDs appear in both), simply concatenating the files won't do.

Anyone had any experience doing such a thing? I have worked quite a bit with BioPerl but have no experience handling annotations. And feeling lazy this time -- would rather find a tool that already exists than roll my own.

Thanks much!!
Bob

[kmcarr]

Do the annotations really need to be interwoven? I know that it is convention but as far as I know it is not a requirement of the GFF. I think it would only matter if the particular software using the file as input requires it.

[BobFreemanMA]

Not sure what the spec says. I tried to concatenate the two files and use that, but the program I submitted the concatenated file to barfed with the error 'duplicate gene id' (or something like that). So i surmised that it expected to see the gene id defined only once, indicating that the data needs to be interleaved.

I guess I could try and take a shortcut -- find all those gene ids that are 'duplicates', remove those 'gene' lines from the 2nd file, and concatenate. Probably about 10 minutes of work...

-Bob

[JohnN]

How about BEDtools 'intersectbed'. It works with gff.