Page 1 of 1

Same gene present on two different fosmids

Posted: Thu Mar 21, 2013 8:04 pm
by pgupta
Hi

We are working on this gene called olf413 and it has 2 identical isoforms.Two projects fosmid 1770P16 and fosmid_2594O08 have the same gene (olf413-PB and olf-PC for one and olf413-RB and olf413-RC for the other).

One project gives blast alignments that do not match up with the browsers software and RNA seq information while the other project does align exons on the browser but the frame for exons keeps shifting from negtative to positive for individual exons.

For the project where blast alignments do not match with the other information on the browser we screened manually and found the exon co-ordinates.
The gene model checker gives a pass for all exons in the checklist but generates a dotplot which looks like this:



Is the dotplot acceptable.We tried shifting the splice sites but this is the best plot we get.
Also do we then annotate the same gene for two different fosmids?



Regards
Paromita
fosmid_2594O08.png
fosmid_2594O08.png (24.06 KiB) Viewed 3477 times

Re: Same gene present on two different fosmids

Posted: Thu Mar 21, 2013 10:42 pm
by wleung
Because the fosmid insert corresponds to a random portion of the D. ananassae genome, the fosmid might end in the middle of a gene. Because the gene olf413 is very highly conserved in D. ananassae, we can map all the CDS's against the D. ananassae 3L assembly using BLAT on the GEP UCSC Genome Browser.

The BLAT results shows that the first two CDS's of olf413 (2_701_0 and 3_701_0) are found in the fosmid_1770P16. The rest of the CDS's are found in fosmid_2594O08. The last CDS (12_701_0) is also found in the overlapping fosmid 2422M21:
BLAT_olf413_D_ananassae.png
BLAT_olf413_D_ananassae.png (103.04 KiB) Viewed 3472 times
Consequently, we should annotate the first two CDS's of olf413 in fosmid_1770P16 as a partial gene missing the 3' end. Similarly, we would annotate the remaining CDS's of olf413 in fosmid_2594O08 as a partial gene missing the 5' end.

You can check partial gene models using the Gene Model Checker by selecting Partial under the "Completeness of Gene Model Translation" field.


> while the other project does align exons on the browser but the frame for exons keeps shifting from negtative to positive for individual exons.

In general, we expect a gene to be in the same orientation and the alignment blocks for the exons should be collinear. Hence BLAST hits which have orientation that differ from the more highly conserved CDS's are likely to be spurious.


>Is the dotplot acceptable?

In addition to removing the first two CDS's from your model, I would also suggest checking the region with a large vertical gap in the dot plot. In particular, the dot plot suggests there are two novel exons in the submitted model compared to the putative D. melanogaster ortholog. The vertical gap in the dot plot (and the BLAT alignment results) suggest there might be an alternate splice donor site for CDS 4 that would account for the missing amino acids and that these amino acids are conserved between D. melanogaster and D. ananassae.

>Also do we then annotate the same gene for two different fosmids?

Yes, please annotate all the genes in your project irrespective of whether the gene appears in other fosmids. In some cases, a project might contain only part of a gene and the gene could span multiple fosmids.