Comment by mbreese | Hacker News Reader

Comment by mbreese | Hacker News Reader

Parent Full thread

No, it does give you the sequence… what it doesn’t give you is the context. The WGS data we’re talking about here only give you two 150bp fragments. You know these fragments are 100-300bp away from each other, but that’s it.

There are other techniques for getting more of the context, but that’s not the WGS we are talking about here.

It’s like copying pages from a book, but without page numbers. And some pages might be copied more than once. Or not at all (Poisson statistics FTW!). You can get fragments of stories, and can probably get the gist of the plot. But this might not give you an exact copy of the full book, in the right order.

To keep going with the analogy — what you’re asking for (telomeres and centromeres) would be the equivalent of placing every “this page left intentionally blank” page in the right place. Even though, they are all quite similar. You will be able to “copy” (sequence) each of these pages, but putting them in the right order is still difficult.

You sum up exactly the problem: the WSG technique does not give the whole global sequence of bases, from telomere to telomere through all of the centromere and for each chromosome branch.

This is what I would like to see as an affordable product: get the whole sequence, so that ultimately one could theoretically re-synthesize their own chromosomes in the future. And get it now that we are still young and there is not too much damage in the DNA, information to be stored for later.

One point in what you wrote: the blank pages (introns) can be really important (like for gene expression regulation, through chromatine folding, or promoters of transcription), we just can't be sure, and that's why it is paramount (when thinking of chromosomes synthesis applications) to have it all.

About the specifics of our discussion: Again, ambiguity! Because we're saying the same thing.

When I say it does not give the sequence, I mean the whole sequence (not only local subsequences), which means including the context.

(Although the notion of context is actually meaningful only if we are talking about local subsequences, as the global sequence is the context (or includes the context, if we reduce "context" to mean the reference global sequence against which the subseqs are aligned)).

Of course with your meaning you are right.

> One point in what you wrote: the blank pages (introns) can be really important

Yes it is -- which is why you get that data with WGS. Seriously, the only sequence that is really difficult to get is around centromeres and telomeres (and some smaller repetitive regions). The other issues that we have with germline sequence is when there are novel structural rearrangements, but with enough depth (or a slightly different technology), we can see those too.

> Because we're saying the same thing

I don't think we are... you keep misusing some of the key terms here. Those "blank pages" weren't introns, they were the ends of the chromosomes (telomere) and the middle (centromere). Those are hard and repetitive. The rest -- the parts you'd really care about if you wanted to do some late life genetic replacement -- are largely covered.

For the context, we use a reference genome. We then look to see where the smaller subsequences align to the the reference genome. This isn't a wet-lab technique, but a dry-lab technique. With enough copies of the smaller sequences (coverage), we can establish what the underlying "full" sequence is. Given sufficient depth, this is a solved problem at this point. The only issues are depth and cost (of sequencing and processing/storage). I left this part out earlier because it wasn't what you were asking about.

We don't have a technique that you can spit in a tube and get the entire sequence of a chromosome all at once. They are simply too long for that to be practical. We do have technology to get a lot (millions) of shorter reads that we can align back to a reference (the company we are talking about in the post does this). We also have technology to get very long sequences (up to 2Mb), but they aren't as accurate as shorter reads. But, if you merge the two together, it can get you all the data you're really after here, even without a reference genome.

Ok, thank you!

So, this means there is hope for getting the data required for a full DNA restoration:

- bases near the telomeres, difficult to get (i.e. difficult to locate) with standard sequencing, can be sequenced (and located) with long-sequence sequencing (but this kind of GWS cheap service doesn't use long-sequencing right?)

- for bases around centromeres, do we assume that they are surrounded by a sea of nothingness, and that relying on a unique reference genome's centromere is enough? Isn't that risky, are we really sure that individual variations do not matter for centromeres?

My aim is to imagine a future where any defects, including centromeres, could be fixed, if necessary by regenerating whole chromosomes. Using data gathered now and safely stored.

This really is putting the cart before the horse though, we will have whole genome assembly technology long before we understand functional genomics well enough to do the type of restorative genetic engineering that you propose.

Well you're right but the idea of getting all of the DNA information is precisely to be sure to not rely on what we think we know, so that if in the future we discover that previously thought useless parts are meaningful we would already have them in store (gathered long before the DNA gets damaged by age).

Once we have everything, we can apply additional knowledge to this data as soon as it is discovered and without needing to do more sequencing.

Reply on news.ycombinator.com