[maker-devel] mising UTRs even with transcript evidence

[Childs, Kevin]

Carson,

I am helping a student to annotate a genome.  Despite having provided transcript evidence, there are many cases where the MAKER gene models are missing obvious UTRs that are found in the transcripts.  The link below is for a screenshot with some examples.

https://figshare.com/s/0b29e0b0ac10f868dbe9

In the figure, there are genes at ~148 kbp, ~152 kbp, ~160 kbp, ~170 kbp that are missing UTRs that should have been predicted by the transcripts in the second track.  The annotation is pretty much riddled with this type of behavior.

One mea culpa, the transcript data was provided as a gff file, and that file had all transcripts defined with a source of “bed2gff” instead of “est2genome”.  I feel that I had found that to be important during some past annotation work, but it doesn’t explain why many but not all genes have UTRs when evidence is present.

Thanks.

Kevin

---
Kevin Childs, PhD

Assistant Professor - Fixed Term
Director MSU Genomics Core Facility
Plant Biology Department
Michigan State University

kchilds at msu.edu
517-775-2844 (m)
517-884-6926 (o)

http://childslab.plantbiology.msu.edu
https://rtsf.natsci.msu.edu/genomics/