Bioinformatics Stack Exchange is a question and answer site for researchers, developers, students, teachers, and end users interested in bioinformatics. It only takes a minute to sign up.
Sign up to join this community
I've had my whole genome sequenced (at 30x average coverage) by a lab, and they have provided the raw files to me (BAM, FASTQ, and VCF).
How can I assemble it?
And does assembly provide any further insights that I can't get from the separated fragments?
I'd like to assemble and annotate it, and find insights, but I don't know where to start.
Given you've tagged this as 'human-genome' there is likely little value in performing genome assembly. De-novo assembly of WGS is generally for non-model organisms without a reference genome available, and would require much higher than 30X coverage (and ideally long reads). By providing the BAM and VCF, they'e already saved you some upstream work unless there are needs to re-process the data from FASTQ
No, assembling won't tell you anything new. When working with organisms whose reference genome is known, all you care about is where your sequences differ from the reference. If you know where your sequences differ, then you can infer the rest, they will be whatever is in the reference.
All you want to do is identify variants, those sites where your genome differs from the reference. That's what the VCF file contains. Bear in mind that the reference isn't some sort of platonic ideal of the perfect human, it's just the collected, pooled DNA of 13 anonymous, apparently healthy individuals. You can expect anywhere from about 3 to 5 million differences in your own genome, the overwhelming majority of which will be completely innocuous.
Since you already have your VCF file, your list of variants, that's all you need to investigate. If you really want to, you can make your assembly by taking the reference genome and then adding the variants in your VCF file, but it won't be of much practical use.
To add the "do not assemble your genome" answers. There might be specific circumstances when you would actually like to do that. For example, if the 30x you got are HiFi long reads and have a reason to suspect there are some interesting large rearrangements in your genome (rare disease?), it might be informative to actually assemble the genome and compare it to the reference to find these rearrangements.
Of course there is also plenty of possibilities in between (still assuming long reads), for example to map your sequencing data to a pan-genome graph or simply use a mapper that is aware of genomic rearrangements (like ngmlr) and call the structural variants.
But in the end it boils down to two aspects you did not mention in your question: What would you like to learn about your genome?
