sakai

Well, the troubling thread here is that that may not be posssible using the amount of blood Theranos collects. Then the question becomes not only why is their valuation so high (which the public doesn't need to care too much about) but "how well do these traditional tests work with Theranos samples?"

That the answer could be "not very well" (and we haven't heard otherwise) is what's most troubling to me.

Hey – working on a startup in the space and am very interested in talking with somebody working in infection control. If it's of interest, please send me an email (in my profile).

And would love to hear your plans too!

Yes, and you've promoted Catalyze.io several times on Twitter in the past month. So let's not feign too much outrage please.

OP here. This was just a little module that I found very useful in quickly exploring large, unfamiliar SQL databases and their associated business logic.

It's fairly rudimentary feature-wise, but has worked well for me. Happy to answer any questions.

Ah – well to clarify, we're performing exact match lookups on individual k-mers (k-length genetic strings), and then computing a classification result from those exact searches.

Yes, we're building out the features to make it easy to use with your own corpus. I'll follow up by email, but can you clarify what species identification issues you're seeing? (Would like to double check and get it fixed asap if there's a problem).

Thx!

That was an attempt to say (without resorting to Big Oh notation) that the data structures and algorithms we're working on provide O(1) lookups that take a fixed amount of time regardless of the reference / dataset size. Does that make more sense?

Hope it was clear that you can test it out with some small FASTA or FASTQ files now (just drag onto the upload icon).

And yes, we'd be happy to discuss your use case and get you set up with something. Just shoot me a note at nick at helix dot io and we can follow up off thread.

In principle, some of them can be yes. Though genetic reference datasets tend to be much larger than, e.g., codebases -- and therefore require different approaches.

One of the core things we're working on is a data structure for more efficient, constant-time sequence lookups, and we've definitely thought of some non-bioinformatics use cases as well. Happy to answer any more specific questions by email too (in my profile)!

cfontes – Yes, that's definitely another use case of the underlying technology, though please note that the database loaded for the demo only include archaeal, bacterial and viral genomes (so you'd have a very hard time getting a hit on sugarcane sequences!).

(This is nkrumm's and my project)

Honest suggestion: Change the tagline to "less than 1000 lines" or exclude the GUI or whatever, but don't just not include line breaks anywhere.

You have 7 very sensible, and indeed admirable, style guidelines and while no line breaks only clearly violates one of them (PEP8), it's not really in the spirit of the others either. I opened gui.py and immediately was overwhelmed, rather than compelled to dive in and digest your small codebase.

Got it. Apparently I tried with a very short sentence! Thanks!

Very cool.

One question – I noticed that you're doing some validation of the inputs (i.e., I can't seem to submit a sentence without a period) – could you explain the reasoning for this?

On the one hand, it seems fairly constraining (and probably comes with a few unfortunate edge cases). On the other, perhaps that's the benefit of the kata philosophy.

Ah ha. Thanks for the heads up!

(Looking over the application again now...)

I find the timing of this and the "New Deal" press releases a bit confusing -- is YC starting to take biotech companies now (like YC S14) or in the next batch?

I'm assuming the New Deal terms are for S14, but was also surprised to see the announcement after the applications were all in.

Can anyone clarify / point me to what I've missed?

Yes! Well said.

It would definitely need to include:

* Web code editor for jobs (JS, Python, Ruby, Perl, Go, etc., etc.)

* Results served over an authenticated API (e.g., servicename.com/api/<username>/<jobname>/files/result.csv would get you the latest result.csv file generated by your script, .../<jobname>/20140419/files/result.csv would get you yesterday's, etc.)

* Jobs could probably be both scheduled (i.e., cron-like) or triggered via a webhook

A simple service for setting up and running jobs/workers without having to run a server.

Ideally, it would have the following features:

* Pay once, run forever (pay for the job up front and never again -- no recurring billing to worry about)

* Configure once, run forever (use Docker/LXC in the background to allow custom environments and absolve the user of the dependency headaches that can arise when running multiple jobs on a single machine)

* Easy to use

I've been casually working on this as it's a pain point I've experienced numerous times (e.g., running a daily job that should cost ~50 cents per month, which is substantially below any available VM price).

Would anybody use this? Other thoughts?

It is really awesome, though I'm kind of bummed that they held off on announcing this (or didn't otherwise finalize the details) until after the summer applications were due. As someone working on some bioinformatics tech, that definitely would have persuaded me to apply.*

* And not primarily because of Elizabeth – who I'm sure is fantastic – but more broadly because of what it signals with regards to YC's focus and energy.

If you haven't seen it already, check out Cryptol from Gallois: http://corp.galois.com/cryptol/

It's a crypto DSL that I believe is implemented in Haskell (it compiles to Haskell, C, C++ and a few others).

Aside: would be very interested in connecting about a metagenomics project I've been working on. My email is in my profile if you want to reach out.

16S reads should be fine for basic characterization. The bioinformatics tools for sequence classification of short reads from a full sequencing run have also improved dramatically recently. Especially in terms of speed (100-1000X improvements).

(Plug) I just launched a site last week that demos this and links to a few other academic works in the field (check the FAQ): https://helix.io

OP here – this is a tool we built for doing really fast metagenomic classification. Happy to answer any questions if folks have them.