Is whole genome pool-sequencing a reduced representation sequencing technique?

Question

[beausoleilmo]

I was reading a paper and it seems that they compare whole genome pool-sequencing with other technique saying that the the cost of reduced representation sequencing technique by reducing the amount of DNA to be sequenced.

In contrast to the whole-genome approach of Pool-seq, the cost savings of these alternative approaches are achieved by reducing the representation of the genome in the sequence data

Reference: Schlötterer, C., R. Tobler, R. Kofler, and V. Nolte. 2014. Sequencing pools of individuals — mining genome-wide polymorphism data without big funding. Nature Publishing Group 15:749–763.

But in the end, whole genome pool-sequencing _is_ a reduced representation sequencing technique because you don't actually read the whole genome. You just don't remove some fragments based on size. Is that correct? Or am I understanding the terminology badly?

Answer 1

[h.mon]

By pooling, you reduce coverage per individual. - I think 1x coverage is sufficient. But you don't reduce the genome, you sample the whole genome. The total cost is higher than genome-reducing techniques, but the cost per marker is lower.

[beausoleilmo]

What I'm usually seen is a 50X coverage for pool-seq on your reads. But since you don't size select, you just read different fragment length. Thus, you are not sequencing the whole genome and are therefor doing a reduced representation sequencing technique. I that correct?

[h.mon]

Forget my 1x coverage, it concerns skimGBS, not pool-seq.

From the abstract of the paper you cited:

Whole-genome sequencing of pools of individuals (Pool-seq) provides a cost-effective alternative to sequencing individuals separately.

Pool-seq is shotgun whole genome sequencing (thus not reduced representation), but instead of just one individual, it is a pooled population of individuals.

But since you don't size select, you just read different fragment length.

I didn't understand your comment above.

[beausoleilmo]

For example, in RADseq, there is a size selection step so that there is "loss" of sequences that will never be sequenced (too large or too small).

From my understanding, if you use illumina sequencing when the genome will be randomly sheared, it's going to be only parts of the genome that will be read (only the end of the sequences that will have been sheared in the "shotgun" process). That would mean that whole-genome approach of Pool-seq using illumina is a reduced representation sequencing technique. So saying that you are doing whole-genome approach of Pool-seq doesn't mean per se that it's the whole genome that is going to be sequenced.

[h.mon]

All Illumina protocols involve a size-selection - size selection is not the reason RADseq is a "reduced representation" technique. Illumina protocols involve a size selection so as to avoid too short fragments (where you would sequence mostly adapters), and to avoid too long fragments, which could interfere with the cluster formation, and thus lower sequencing quality.

RADseq "reduces" by linking adapters to restriction-site cuts, read this overview about RADseq (from which the fragment bellow is quoted):

By attaching customized adapters to the cut ends of the fragments, it’s possible to ensure that a high-throughput sequencing system will only sequence from a specific subset of the fragments.

The differences in coverage due to random shearing (leaving aside the fact that the shearing is not truly random) is not by design, it is an unintended consequence of whole genome shotgun sequencing and associated techniques to fragment the DNA. According to your logic, all genome assembled to date are "reduced representation", as they were assembled using whole genome shotgun - luckily, nanopore sequencing will fix this, as we will be able to sequence whole chromosomes.