Efficient fetching of alternative sequences?

Science
1

Hi, I’m currently searching for an efficient way to fetch alternative sequences from range queries on genomic variants.

Therefore, it is necessary to fetch for one individual all the variants in a certain genomic region.

Is Hail suitable for this kind of query pattern?
I’d like to use this for 10,000s of samples.

[EDIT for better explanation:]
For example, some individual might have on position 2 an A->G mutation and at position 4 a deletion of three base pairs.

Reference sequence: TATCCCGGG
Alternative sequence in range [1…6]: TGTGGG

To calculate this alternative sequence, it is necessary to fetch for one individual all the variants in a certain genomic region.

2

Sorry, you’re using some lingo I don’t totally understand:

fetch alternative sequences

Can you give an example of what you would want to get out of this query on VCF-like data?

3

For example, some individual might have on position 2 an A->G mutation and at position 4 a deletion of three bases.

Reference sequence: TATCCCGGG
Alternative sequence in range [1…6]: TGTGGG

To calculate this alternative sequence, it is necessary to fetch for one individual all the variants in a certain genomic region.

4

Wouldn’t this be represented as two different non-overlapping variants, one at position 2, one at position 4?

Do you have a precise definition of an “alternative sequence”?

I still don’t see what you’d want to do with VCF-like data here, sorry.

5

Yes. Given some individual, I need to know all its variants in a certain region.

Variant == difference to reference genome, i.e. if I know the reference genome + the variants in the region, I can calculate the individual’s DNA sequence in the region. That’s the alternative sequence I try to obtain for each individual.

So, in short I need to execute range queries for single individuals. Does this make sense to you?

6

Yes, got it. Hail’s typical representation of VCF-like data stores a variant-major matrix, which does not make it easy to do rapid single-sample lookups. If you want to execute a query like this for all samples, Hail can probably do that efficiently.

1 Like
7

Thanks for explaining. Yes, I’d like to do this for all samples in the end, but I’d need to get this batch-wise.