bioinf / Sibelia

Genome comparison via de Bruijn graph. To get the latest stable version, please visit our site.
http://bioinf.spbau.ru/sibelia
Other
64 stars 17 forks source link

Preventing self search? #175

Open ghost opened 7 years ago

ghost commented 7 years ago

Hello,

Is there an option to prevent Sibelia to find synteny block inside one of the 2 input genomes? For example, when I compare genome_1 and genome_2, Sibelia will not only find the syntenic regions of the 2 genomes but will also find the duplications inside each genome.

I think it would be interesting to have an option to avoid that, so that we could just get the synteny blocks involving the 2 genomes. The "sharedonly" option doesn't give me what I want, because if the genome is polyploid then it returns empty files.

iminkin commented 7 years ago

What should the output be if genomes are polyploid? For example, suppose you have two genomes:

G1: +A +A +B +B G2: +A +B +A +B

Where A and B ary synteny blocks. So what would be the desired output in this case?

ghost commented 7 years ago

It would be, if we just consider the first "A" of the G1 genome

G1 +A and G2 +A G1 +A and G2 +B etc ... And never G1 + A and G1 +B

If I understood correctly your example.

iminkin commented 7 years ago

This is tricky because it is not clear why we should choose the first appearance of "A" as the "main" one, and not the second. There is no ready-made option to impose this behaviour in Sibelia. You can write a simple postprocessing script that takes ouput of Sibelia and "resolves" duplications the way you like.

ghost commented 7 years ago

Yes indeed, this is the approach I have undertaken. I am working on a very unusual organism for which no genomics tool is really well adapted (the bdelloid rotifers, which are degenerate polyploid with no homologuous chromosomes, which means I have to work with all allelic pairs or paralogs present. There is also no way to distinguish an allelic pair from a paralog). Sibelia is nevertheless the best so far, I really like your program.

iminkin commented 7 years ago

That is cool :) I am glad that you found Sibelia useful.