Nanopore sequencing, Shasta toolkit enable de novo assembly of 11 human genomes | Hacker News Reader

Nanopore sequencing, Shasta toolkit enable de novo assembly of 11 human genomes

nature.com

32 pointsbookofjoe17 comments

Oxford Nanopore (makes the sequencer used here) is one of my favorite companies. I think we'll see their work start to change how we do diagnostics in the next 5 years, especially with their automated library preparation product [1].

I was blown away how possible billion-base DNA sequencing is in non-lab environments when we bought one of these for fun [2].

[1] https://youtube.com/watch?v=x-z3MFkpvzE&t=1m20s

[2] https://abarry.org/dna-sequencing-in-our-extra-bedroom/

That's kind of epic that you managed to pull that off in a home setup. How exactly did you manage to run the gel and all? Did you buy a gelbox/agarose/dye etc too?

ONT have a ‘rapid’ kit that doesn’t require standard DNA lab equipment.

1. https://store.nanoporetech.com/us/sample-prep/rapid-sequenci...

The linked article in the patent comment talks about running a gel to confirm extraction which is what I was referring to

We used the rapid kit and also a gelbox / agarose.

Did you end up buying everything? I'd be curious to know how much it cost overall if you'd be willing to share. While most of my time has been in well stocked academic labs or shared startup lab space, I did spend some time working out of a bio hacker space and I'm always curious how cost efficiently people can wrangle together enough otherwise expensive kit to run an experiment in an odd setup. Did you end up with a lot of extra agarose etc or were you able to resell the leftovers and the used kit? Super cool stuff.

In total it was about $3k. So not cheap, but also not a real lab budget by any means.

Thanks! I skimmed but noticed that they used the rapid kit - sorry!

Highly entertaining write up! Hats off to you for attempting it (successfully!).

Shasta produced a complete haploid human genome assembly in under 6 hours.

I found out the record is 100 minutes. https://www.biorxiv.org/content/10.1101/705616v1.full

The improvement could be that it is done on a single commercial compute node.

It's a bit frustrating (and concerning?) that their data repository [1] seems incomplete; the most important seqences (GM12878 / HG001) are missing from everything but the raw data directory. This is the sample that most laboratories would likely have readily available for comparison to their results, but for some reason they have omitted to deposit their final results for it (unless I'm missing something ...).

[1] https://s3-us-west-2.amazonaws.com/human-pangenomics/index.h...

Is it in any of the supplementary information or in any of the figure source data links in the paper? They do link in addition to the AWS open data this link (https://www.ebi.ac.uk/ena/data/search?query=PRJEB37264) - did you look through that? Other than that now idea - seems like a bit of an oversight from nature if they wouldn't make sure important sequences were readily available, you'd imagine they'd be important for peer review as well prior to publication.

Look at their "sample selection" section. "The goal of sample selection was to select a set of individuals that collectively captured the maximum amount of weighted allelic diversity53." And also says "We also added the three Ashkenazim Trio samples and the Puerto Rican individual (HG00733). These four samples were added for the purposes of comparison with other studies that are using them."

Four genomes make a comprehensive evaluation set, don't think HG001 is "absolutely" needed.

a.k.a. days of gene-edited humans.

This is a faster sequencing technology, not a faster gene editing technology.

And the days of gene edited humans have been here for a while now. Gene therapy trials have been running for decades, fda approved in 2017 ish, etc

Faster sequencing will enable faster gene editing turnarounds by faster mutation detection.

Jurassic park was the first thing I thought of. But it's certainly one tool in the toolbox for that.

Reply on news.ycombinator.com