Tool to detect gene replacement in bacteria with NGS

Question

[kmyers2]

We are re-sequencing several strains of our bacteria (Rhodobacter sphaeroides strains) to identify mutations and differences from the reference R. sphaeroides genome we have. I've used breseq and GATK to identify SNPs and indels with success in the past. However, we are sequencing older strains in which we have replaced a gene of interest with an antibiotic gene. When I process these samples with breseq or GATK it does not list the gene of interest as being mutated, but does identify other SNPs and indels.

I have verified that the gene of interest lacks coverage by plotting the BAM files, but this can be laborious for many strains.

Is there a good tool that can identify gene replacement instances such as this? We are likely going to be sequencing many, many more strains in the future and I would like to have a way to detect these without having to manually look at the coverage for the region in the genome.

Answer 1

[LChart]

I haven't seen published tools that do this; but I have written quite similar tools for verifying knock-ins and knock-outs. Typically I will compute the ratio of median coverage over the knock-out gene (the annotation .gff can be used to define the locations) to the median coverage over the genome, and expect it to be <10%. For knock-ins I typically create a new fasta by concatenating the original genome with the construct that was inserted (typically it's a random insert, so I'll have it as a separate contig); and compute the same ratio, expecting it to be >80%.

[kmyers2]

Thanks! That's a really clever way to go about it and detect these kind of knock-ins/knock-outs.

[LChart]

Not quite as elegant as just using PCR (which is also performed); but the genome is being sequenced to identify off-target effects, anyway; so these ratios are used as a secondary confirmation.