[maker-devel] Maker-derived CDS GFF3 phase column

[Rob Syme]

Hi all

By my reading of the GFF3 spec (http://sequenceontology.org/resources/gff3.html), I'm getting gff3 from Maker that has odd data in the phase column.

For example, see some example Maker output at https://gist.github.com/robsyme/5617399

There are two exons, 5617 <- 5737 and 5793 <- 5953 with phases 0 and 2, respectively. Both exons are in the reverse strand.

From the spec, phase indicates "the number of bases that should be removed from the beginning of this feature to reach the first base of the next codon", and for "reverse strand features, phase is counted from the end field". 

In the case of the 3' exon (5793 <- 5953), the end field (the 5th column) is 5953. 

The base at the end field is the first base of the translated CDS, so there should be no bases removed "to reach the first base of the next codon". I suggest that this phase should be 0, not 2.

There is an illustration of the feature at http://i.imgur.com/DKLxnSf.png.

The output gff3 is correct if "the number of bases that should be removed from the beginning of this feature to reach the first base of the next codon" is measured from the 'left-hand' end of this feature (the start field) rather than the end field.

Has anybody else ran into this problem or am I misreading the gff3 spec?

Rob Syme

PhD Student

Curtin University

[Carson Holt]

It looks like the phase was calculated from the wrong strand orientation.  I believe I have corrected this now.  I'm checking a few more things, but I'll have 2.28 as the latest release likely tomorrow with the cumulative bug fixes since the last release.

Thanks,

Carson

_______________________________________________ maker-devel mailing list maker...@box290.bluehost.com http://box290.bluehost.com/mailman/listinfo/maker-devel_yandell-lab.org

[Rob Syme]

Fantastic. I thought that might have been the problem.

Looking forward to 2.28. Thanks!

Rob

[Carson Holt]

Ok.  It's available for download.

--Carson