Help with genome comparison between two strains.

Question

[Andrea Laura]

Hi, I'm new to bioinformatics and I need to compare two bacterial genomes. These genomes come from different strains of the same species. and both genomes are available in NCBI

What I need to do is a genome wide comparison to find the homologous genes in both sequences and what is the name of both locus tags.

In other words I need a table that says Gene A corresponds to Gene B.

I've searched around and found some tools that might help me, for example, microbializer (which is not working) and orthologr but after 10 days of running a power shortage terminated my run and the results I could get were in protein code and not in gene code.

Anyone knows any tool that might help me?

[Jimmy]

Couldn't you BLAST every gene in one strain against the genome of the other strain and vice versa?

[Dunois]

That is basically what a reciprocal best hit (RBH) search is. This is the approach suggested here: Help with genome comparison between two strains. .

Answer 1

[dthorbur]

What about microbializer isn't working? I see their github was updated a year ago. What have you tried to get it working? And what are you running things on? 10 days to compare 2 similar genomes seems excessive.

Other tools you could look at include OrthoMCL and OrthoFinder. That's good for genome-wide analyses as they clusters based on sequence similarity. If you want to be more sure that a transcript/protein is the same in both assemblies you could use a hmm based tool like HMMER.

If you also download an annotation file (GFF/GTF), then you can see what proteins/transcripts are part of which genes and which chromosome/contig they are located on.

[Andrea Laura]

As far as I could see microbializer is a web tool and that isn't working. I tried to find how to run it locally and didn't find how to do so. The other program i tried I run it on my personal laptop, thats all i have access to :/

Answer 2

[Roberto]

You can use https://github.com/soedinglab/mmseqs2/wiki#reciprocal-best-hit-using-mmseqs-rbh. For two bacterial genomes you can run it easily on a laptop in a matter of minutes. It's basically like blasting genome vs genome and returns the best batch for each feature.

[Dunois]

This is the way. Something like OrthoFinder is overkill for a single pairwise comparison.