How to create chain file to change genome coordinates from one genome version to another

Question

[iraun]

Hi there,

I'm looking for a 'protocol' or I would like to know the best way to create a chain file in order to get the new coordinates of a new assembly of a given genome. I have an old genome, with a gtf file, and a new genome (only the fasta). I would like to create a new GTF, and I'm thinking about Crossmap, but first, I need this chain file... I have read something about blat if the specie is the same (my case)..: http://genomewiki.ucsc.edu/index.php/Minimal_Steps_For_LiftOver, but my genome is larger than 4Mb!

Anyone knows a pipeline o protocol for doing this?

Thanks in advance,

[Jada]

Hi,

I run into the same problem and Crossmap looks promising.. However, it looks like the chain file is critical..

Have you come up with an alternative strategy to resolve this issue? Any help/guide is more than appreciated.. Thanks..

[fransua]

I just found this: https://github.com/wurmlab/flo haven't tried yet, but it seems to do the job...

[Pierre Lindenbaum]

what about http://genomewiki.ucsc.edu/index.php/Same_species_lift_over_construction ?

[iraun]

Thanks Pierre. Yes, I have already read that link but... "expecting small genomes with chromosome sizes less than 10,000,000 bases each ". What's up if my genome is larger?

[iraun]

Maybe I can change the parameter export targetChunkSize="10000000" (inside the script), up to my genome size but... maybe the rest of variables in the script are set to that size and not a bigger size..

[Pierre Lindenbaum]

the 'targetChunkSize' is not the genomeSize. The UCSC use that script to generate the chain files for the human genome.

[iraun]

Look here, looking the last comment... he says that those scripts are only for small genomes...