Hacker News Reader

Comment by valine | Hacker News Reader

How do we get to the point where this therapy doesn’t require a team of scientists to manufacture? Is there a future where every cancer center has a DNA sequencer and mRNA synthesizer?

View on HN

Googling the cost of DNA sequencing suggests that "human genome sequencing cost decreased from $1 million in 2007, to $1000 in 2014". That's pretty amazing.

It’s been 9.5 years since 2014. Wonder how far down it’s come since then.

The $1,000 'cost' was the actual lab cost previously, we're now at a point where consumers can buy WGS for under $400, so the equivalent supplies/materials cost must be well below that;

https://sequencing.com/products/purchase-kit

With the Novaseq X the cost is approximately $200 a genome. One the current challenges is storing the hundreds of gigabytes of data per genome.

1TB SD cards are a thing for 150USD now? Maybe two for physical redundancy and some Reed-Solomon stuff on each card?

I wonder if that would be cheaper in the long term than the typical NFS plus off site tape backups and/or cloud backup used by the places I have worked. I suspect it would become a logistical nightmare to store and keep track of thousands of SD cards.

I mean we still give DVDs for MRI results here...

Edit: I'd prefer my biostuff data to stay mine and not be saved in multitude cloud piñatas...

It'd be pretty amazing to have something like the optical storage jukeboxes but for SD cards. Expensive, but amazing.

I found the issue about storage quite interesting, went digging and I found this article: https://medium.com/precision-medicine/how-big-is-the-human-g...

Although the article does not mention it anywhere, I thought such repetitive byte sequences could be compressed quite well with simple file compression. Went digging again. I was wrong, compression of FASTQ genome sequence files IS a thing: https://en.wikipedia.org/wiki/FASTQ_format#Compression

It goes far beyond that.

By the time any real analysis is done on these data, they're extremely compressed via other file formats that extract the important signal. Alignment data is stored in BAM/CRAM [1], [2] formats, and variant data in VCF [3]

[1] https://en.wikipedia.org/wiki/Binary_Alignment_Map [2] https://en.wikipedia.org/wiki/CRAM_(file_format) [3] https://en.wikipedia.org/wiki/Variant_Call_Format

Is there a service that burns your results after giving them to you?

I don't need an AI eugenic death squad coming after me in 20 years.

What about an AI-dispatched procreation team coming to milk you for everything you’ve got?

https://nebula.org/whole-genome-sequencing-dna-test/

Depending on options like depth and quality, a hundred to 500 hundred dollars.

Don't forget "Your Insurance Will Totes Pay for It" CEOs like the Epi-pen guy.

500 hundred (50,000) may still be low.

Got mine done at 100x coverage (BGI genomics short read) for about 1000$ a couple years ago.

(The 1000$ figure you're citing is probably benchmarking against 30x coverage, so 1k for 100x is a pretty good price.)

Part of my research involves designing these cancer vaccines. There are different delivery mechanisms, but the need to synthesize medical-grade peptides or mRNA vaccines is one of the major drivers of cost. The sequencing costs are pretty minimal at this point, but analysis of that data by a team of experts is part of that cost as well. This is still largely bespoke stuff that can't quite be standardized because we're still figuring out the best approaches, but there are quite a lot of people working on it!

By doing research like this. Universities prove it's viable, then companys productionize it, sometimes with incentives from the government

Also, how do we get there without effectively creating bioweapons lab capable of immense damage in the wrong hands?

Reply on news.ycombinator.com