Comment by hirenj | Hacker News Reader

Comment by hirenj | Hacker News Reader

First thoughts reading up on this pretty cool bit of detective work: Since this is a deficiency in production of Pngase F, I'm not sure a simple injection of Pngase F is going to work. I'd guess you'd need to target Pngase F into the ER to kick start the proper clean-up process, and the human form (http://www.uniprot.org/uniprot/Q96IV0) is different to the recombinant form that you can get synthesised (http://www.uniprot.org/uniprot/P21163). I'm not really sure how native Pngase-F is regulated!

That said - I'm about to start work on very similar work now (also in the area of glycobiology). With dirt cheap exome sequencing, we're going to get a whole bunch of really interesting leads from the data. This means that the follow-up research into the mechanism behind the action of the gene can be more likely to yield results.

Right now, I see the bottle neck in this whole process being the actual experimental analysis of these mutations. Once we solve how to scale up this hard work successfully, we can start looking at curing these incredibly rare diseases.

Wow! This is fascinating.

Thank you so much for the pointers!

Are you a grad student/postdoc/faculty?

I'm afraid it isn't the most in-depth of a write-up!

I'm a postdoc computational biologist at the Copenhagen Center for Glycomics, so you could say this is somewhat up my alley ;)

Feel free to fire off an email (in the profile) if you've got any questions, and I can try to help out.

In what way is experimental analysis of such mutations a bottleneck?

Well, once you pinpoint a potential mutation on a gene - that gives you a starting point into the pathway that could be possibly producing this particular disease. From there, it's sort of like clicking outwards from a wikipedia page, except instead of wiki pages, its related genes. Every one you go to, you read and find out what people know about it. You then start to form a picture in your head as to what the mechanisms of action are for this particular gene - what it does and how it is regulated.

From there, you build your model for how the gene works, and do some knock-outs / knock-ins, test it in various cell lines, verify the kinetics, try to find out the 3D structure. Really, anything you can do to get a handle on how this actually works in the system. It's quite normal for people to spend their entire PhD studying the mechanism of action for a single gene!

What can be done to speed this process up? What parts can be automated? Is the process for every gene unique or can some steps be factored out?

The common factors in this are having smart people looking at it, unfortunately. We're working on various ideas to become much more efficient at it, but it's a really hard slog.

Reply on news.ycombinator.com