Question regarding CDS frame

[John Martinson]

In EVidence Modeler, is the value (0,1, or 2) in the frame field of a CDS record in a gff3 file relative to the entire contig/scaffold/chromosome, at least for the position of the beginning of the coding sequence? For example, lets say my scaffold is CCCCATGCACTAGCCCC. Total length is 17. Start codon (beginning of CDS) begins at base 5. From what I can tell if Augustus was used to generate a prediction that started at the ATG, it would report a frame of 0 (it appears to me that Augustus, at least when reporting a CDS that commences with the start codon, always says it is in frame "0"). With Genemark (which uses a 1/2/3 frame scale), I think it would report the frame relative to where it is on the whole contig, so it would report a frame of "2" for the CDS beginning at the ATG. I think Snap also reports the frame relative to position on the entire contig, only on a 0/1/2 scale. Regardless of whether I'm correct about all that or not, what does EVidence Modeler expect? I am parsing results from Augustus, Snap, and Genemark into an appropriately formatted GFF3 file to use as input to EVM and I want to make sure I get the correct values in the frame field for the CDS records.

Thanks.  

[Brian Haas]

Hi John,

EVM doesn't take into consideration the frame values provided by other programs and instead reevaluates them directly based on how the coding region looks in the genome based on the exon coordinates.  So, no worries there.

best,

~b

On Fri, Feb 22, 2019 at 9:11 PM 'John Martinson' via EVidenceModeler-users <evidencemo...@googlegroups.com> wrote:

In EVidence Modeler, is the value (0,1, or 2) in the frame field of a CDS record in a gff3 file relative to the entire contig/scaffold/chromosome, at least for the position of the beginning of the coding sequence? For example, lets say my scaffold is CCCCATGCACTAGCCCC. Total length is 17. Start codon (beginning of CDS) begins at base 5. From what I can tell if Augustus was used to generate a prediction that started at the ATG, it would report a frame of 0 (it appears to me that Augustus, at least when reporting a CDS that commences with the start codon, always says it is in frame "0"). With Genemark (which uses a 1/2/3 frame scale), I think it would report the frame relative to where it is on the whole contig, so it would report a frame of "2" for the CDS beginning at the ATG. I think Snap also reports the frame relative to position on the entire contig, only on a 0/1/2 scale. Regardless of whether I'm correct about all that or not, what does EVidence Modeler expect? I am parsing results from Augustus, Snap, and Genemark into an appropriately formatted GFF3 file to use as input to EVM and I want to make sure I get the correct values in the frame field for the CDS records.

Thanks.  

--
You received this message because you are subscribed to the Google Groups "EVidenceModeler-users" group.
To unsubscribe from this group and stop receiving emails from it, send an email to evidencemodeler-...@googlegroups.com.
To post to this group, send email to evidencemo...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/evidencemodeler-users/6dee2fcf-e52d-48c5-9c41-fbdc75ef38d5%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

--

--
Brian J. Haas
The Broad Institute
http://broadinstitute.org/~bhaas

 

[John Martinson]

Brian,

Thanks for the quick response. Next time I'll remember to ask first before spending hours trying to figure something like that out blindly. That said, the GlimmerHMM example in the "Preparing inputs for EVM" section of the web page does show numeric values in that field for CDS records, which is why I thought I needed to figure it out. You might consider changing that or making a note about it there so others don't make the same erroneous assumption I did.

Thanks again,

John

Sent from Yahoo Mail on Android

--

[Brian Haas]

Sure thing.  Thanks for mentioning this.   I need to move the documentation over to the wiki format and when I do I'll add a note about this.

best,

~b

To view this discussion on the web visit https://groups.google.com/d/msgid/evidencemodeler-users/1764563177.3778070.1550892549436%40mail.yahoo.com.

For more options, visit https://groups.google.com/d/optout.