SNP calling with de novo assembly

Question

[qinghong]

I am new to SNP. I am looking for some help. My PI suggested me to call snps with de novo assembly . I have 100 bacteria draft assembly (each assembly has multiplex contigs). Should I use mummer (nucmer) to align each draft genome to the reference, call the SNP. After that combine all the snps into one file? I checked the mummer, It seems mummer could not align multiple draft assembly to one reference. If I am right, what tool I could use to combine the SNPs from different assembly in to one file? Thank you very much!

Rachel

Answer 1

[Vitis]

Would you describe your objectives a bit more clearly? Are you trying to conduct comparative genomics across the strains, or simply calling variants among these strains? If your strains are closely related, you may choose one strain as a reference, and use minimap2 to align contigs of other strains agains the reference. I think the resulted BAM files could be subjected to a standard variant calling process. The decision of choosing which strain as the reference is an interesting question itself. Sometimes this can be guided by a phylogenetic tree generated with multiple-genes so you can pick a "representative" strain.

Mummer only works for pairwise comparisons, not multiple sequence alignments. If the strains are relatively distantly related, which would have much more changes and differences across strain than simple SNPs, you may need other comparative genomics tools to align the contigs and call the changes.