1
$\begingroup$

I was reading the manual of MCScanX and found that besides a all-vs-all protein blast file the tool takes a gtf file with four columns only: scaffold name, protein name, from, to.

As the tool is not taking the strand column, it appears to me that probably the most popular tool for genome collinearity analysis is disregarding orientation of genes. Do I get it right? Is there an analogical tool that would perform the collinearity analysis with respect to orientation of genes?

-- edit --

Now I am almost entirely sure that it MCScanX is not taking orientation into account since I found collinear blocks where some of the gene homologs were on the same strand and some of them two different strands (I do hesitate to call them co-"linear" now). So the updated question would be: Is there a good MCScanX alternative that would take orientation of genes into account?

I am not sure if I will improve the question with an example, but i will try anyway. Let's say that A B C D was detected as collinear with A' B' C' D'. The orientation of genes is + + - + for A B C D and + - - - for A' B' C' D'. For sure, these are synteny blocks. The order of genes is the same, but are they collinear? How can I search just for blocks that would be same (or completely revers) orientation?

$\endgroup$
4
  • 1
    $\begingroup$ Sounds like you could just do logic on the from, to columns if they correctly have the from > to if the gene is in the reverse orientation as the reference. $\endgroup$ – conchoecia Oct 29 '18 at 16:51
  • $\begingroup$ Theoretically it would be possible to encode the orientation information like this, but that would be against gtf specification that has the leftmost and the rightmost positions. $\endgroup$ – Kamil S Jaron Oct 30 '18 at 9:28
  • $\begingroup$ Couldn't that be taken care after the "co-linear" analysis? In your example what would you consider co-linear (B and D with their pairs B' and D')? Sorry it is not my fields but seems interesting $\endgroup$ – llrs Oct 31 '18 at 9:16
  • $\begingroup$ @Llopis Yep, I have ended up in post processing the MCScanX output - I need to post an answer. In the example I made there are two different collinear blocks A+ C- C'- A'+ (reverse duplication) and B+ D+ D-' B'- (a palindrome). Assuming that we allow gaps in between genes. $\endgroup$ – Kamil S Jaron Oct 31 '18 at 18:01

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.