DIYbio-esque App wins 1st at YC-Hacks

[Will Sutton]

Athelas, an iphone app to do hematology (blood imaging) won first place in a Y-Combinator hackathon this weekend. It strikes me as very DIY-esque, doing something complicated and expensive on a minimum budget.

Over in the discussion on H/N, I'm seeing a lot of poo-poo-ing of the feasibility by actual technicians.

What do you guys think: Can good ML on cheap devices increasingly replace lab tests? Also, is this a harbinger of more buy-in for DIYbio concepts from the mainstream venture community?

[Bryan Bishop]

On Mon, Aug 4, 2014 at 10:06 AM, Will Sutton <wsut...@gmail.com> wrote:

Also, is this a harbinger of more buy-in for DIYbio concepts from the mainstream venture community?

That sounds like regular old venture backed biotech to me. There's nothing wrong with venture financing in that sense, but I wouldn't really call that do-it-yourself if you're raising VC. I mean, you're welcome to steal the marketing I guess, but then everyone hates you and you sort of slowly kill it.

- Bryan
http://heybryan.org/
1 512 203 0507

[Nathan McCorkle]

The images in the first link you posted are quite crappy IMO. Garbage
in, garbage out.

> --
> -- You received this message because you are subscribed to the Google Groups
> DIYbio group. To post to this group, send email to diy...@googlegroups.com.
> To unsubscribe from this group, send email to
> diybio+un...@googlegroups.com. For more options, visit this group at
> https://groups.google.com/d/forum/diybio?hl=en
> Learn more at www.diybio.org
> ---
> You received this message because you are subscribed to the Google Groups
> "DIYbio" group.
> To unsubscribe from this group and stop receiving emails from it, send an
> email to diybio+un...@googlegroups.com.
> To post to this group, send email to diy...@googlegroups.com.
> Visit this group at http://groups.google.com/group/diybio.
> To view this discussion on the web visit
> https://groups.google.com/d/msgid/diybio/0876a71f-c542-4821-a054-bbe2d7197421%40googlegroups.com.
> For more options, visit https://groups.google.com/d/optout.



--
-Nathan

[John Griessen]

On 08/04/2014 11:46 AM, Nathan McCorkle wrote:
> The images in the first link you posted are quite crappy IMO.

Yes, but...
They say, "The attachment magnifies/focuses on the sample by means of a 1mm ball lens."

That's sounding like Van Leeuwenhoek's early microscope lenses, and yes the images look
very so so, but maybe the imaging software can do an edge detect and lose all the out-of-focus-blur
and analyze what's left... Instead of a perfect image, it may be more important to
scan along the whole area under consideration, then change depth of focus, rescan,
repeat-while-necessary to get a 3D map of images, then recombine some of them if needed
for a positive ID of what is floating at that spot. If you're talking about doing White cell count
and red cell count, identifying parasites is not so important. You already know you started
with someone's blood, so all you need to tell is is it a WB cell or a red corpuscle.

Getting a volume correct might be more valuable and maybe a standard count by humans
uses such a small volume squashed under a cover slip that a lancet drop
would give better counts than a big old vial.

> Garbage in, garbage out.
>
> On Mon, Aug 4, 2014 at 8:06 AM, Will Sutton <wsut...@gmail.com> wrote:
>> Athelas, an iphone app to do hematology (blood imaging) won first place in a
>> Y-Combinator hackathon this weekend. It strikes me as very DIY-esque, doing
>> something complicated and expensive on a minimum budget.

That's just tech innovation is all.

>>
>> Over in the discussion on H/N, I'm seeing a lot of poo-poo-ing of the
>> feasibility by actual technicians.
>>
>> What do you guys think: Can good ML on cheap devices increasingly replace
>> lab tests?

What do you mean by ML?
Machine Lab?

Yes, it going to be the answer, but not on cheap generic hardware, on semi secret sauce
MEMS/CMOS fabbed silicon hardware by theranos.com See this article: http://fortune.com/2014/06/12/theranos-blood-holmes/

Also, is this a harbinger of more buy-in for DIYbio concepts from
>> the mainstream venture community?

Like Bryan said, no, the VCs want leverageable ways to offer more for less in high volume
to the masses and delivered by teams of twenty-somethings that will work for stock options
and make them, (the VCs), rich by 100X ROI.

There's nothing DIY about more for less.
More bang for less bucks the driver behind business in general.

DIY is often less for less. Less as in -- with the materials at hand, crude and unpolished.

[Nathan McCorkle]

ML I assumed was machine learning.

Yes I know they used a ball lens, I still think "garbage in garbage out" applies. They said hemotology not cell counts, so I assume you need better data, but I could be wrong.

--
-- You received this message because you are subscribed to the Google Groups DIYbio group. To post to this group, send email to diy...@googlegroups.com. To unsubscribe from this group, send email to diybio+unsubscribe@googlegroups.com. For more options, visit this group at https://groups.google.com/d/forum/diybio?hl=en

Learn more at www.diybio.org
--- You received this message because you are subscribed to the Google Groups "DIYbio" group.

To unsubscribe from this group and stop receiving emails from it, send an email to diybio+unsubscribe@googlegroups.com.

To post to this group, send email to diy...@googlegroups.com.
Visit this group at http://groups.google.com/group/diybio.

To view this discussion on the web visit https://groups.google.com/d/msgid/diybio/53DFCF0D.9090902%40industromatic.com.

[Will Sutton]

On Monday, August 4, 2014 11:13:43 AM UTC-4, Bryan Bishop wrote:

That sounds like regular old venture backed biotech to me. 

It was built by a high school student in 30 hours with a piece of rubber, electric tape, and toilet paper...how does that sound like regular old biotech? I think he gets a small prize, not financing to start a business yet.

What I think is compelling is that it beat all the other social/bitcoin/location app non-sense that people are spending their time on. If there were more success stories like this, I think there'd be more people trying to get involved in bio-based tech.

 

[Marc Dusseiller]

good morning,

seems like it's hard to beat good ole Van Leeuwenhoek.... as some might have seen, i have been fiddling around with cheap DIY microscopes for a couple of years.

http://hackteria.org/category/projects/diy-microscopy/

you can download a nice bilingual instructions here

http://hackteria.org/wp-content/uploads/2013/05/hm3-hackteria-pages.pdf

also i did many tests adding a lens to a smart phone. works very easily and it's possible to get much better pics, than the ones mentioned above. important is to get a stable setup and decent lighting condition. 

http://www.instructables.com/id/10-Smartphone-to-digital-microscope-conversion/

http://www.bioartsociety.fi/workshops/?p=45

you can try lenses from old webcams or laser pointers to start with, or small spherical lenses that you can also buy... or even make yourself leeuwenhoek staaaayle.

http://www.funsci.com/fun3_en/antoni/vlen.htm#3

other options is to hack the lens from a DVD drive to be able to focus electronically.

http://hackteria.org/wiki/DIY_Laser_tweezer,_cell_trap,_oligo_synthesis

the internet is full of instructions and hacks to turn your i-phone into a decent microscope. this young designer just made a nice clip-on... still talking to him to share the design on thingyverse.

http://designerei.ch/?id=26_my-iCros_Mikroskop

i am happy that all channels, hackathons, media are used to spread the message of building DIY microscopes is easy. especially when we think about other places in the world, eg africa.

http://hackteria.org/discourse/hackteria-poc-diagnostics-workshop-kenya/

https://vimeo.com/99143018

we succesfully imaged trypanasomes, a parasite in the blood that causes sleeping sickness.

the claims of the above mentioned "app" to detect malaria "cells"... i am a bit skeptic, also generally pissed of by this post-colonial hi-jacking of the big health challenges of the developing countries to win money in the west during hackathons. malaria is caused by another parasite of the genus plasmodium, especially p. falciparum. but yes, microscopy is the so called gold standard of diagnosing it. new fluorescent staining methods have made a big impact recently, combined with "cheap" LED powered microscopes. but they are still quite expensive.

http://www.finddiagnostics.org/programs/tb/find_activities/led_microscopy.html

but still I think there is a big potential for the DIY microscope, either for a smart phone (which is not low-cost) or using hacked webcam, for malaria blood smear tests and other neglected tropical diseases... or in the end just in public education all over the world. generally an automated and computer vision detection would be great!

my friend akbar has already developed new nice version of the MICAM microscopes and gave them to schools all over indonesia.

http://www.slideshare.net/HLab14/micam-akbar-lifepatch

greets,

marc

[Jonathan Cline]

What do you mean by "DIYbio concepts".  Do you mean, good engineering?   Which by the way, "high school student in 30 hours with a piece of rubber, electric tape, and toilet paper" is not good engineering; good engineering would include a CAD drawing as a model suitable for manufacture.  Also, I thought "DIYbio concepts" should include by default that the solution actually does what it claims to do; rather than what is sometimes posted here as journalist-bait: performance claims made before proof, and re-invention claimed as novel.   Who were the judges of the competition and what is their criteria for 1st?  Vaporware often wins such competitions, so does an impassioned speech (vs. technical merit).  Science and good engineering typically takes longer than 30 hours.  Is it significant that a bio-related project might win such a competition?  Yes.  As part of a general trend starting sometime in the mid 1990's, the hype train has been slowly building for biotech kits and bio-monitoring/measurement, and will continue to build, especially thru newer sensors and connected computing (internet, mobile devices, cloud stuff); in general this is the slow yet inevitable convergence of many technologies long predicted by many hackers, driven by consumer demand for communications technology.  Consumer demand to be able to tweet and netflix anywhere is fueling biotech R&D, in essence.  Is it getting easier to rip off VC money with fake biotech projects?  Yes - it has always been easy - and now it is easier than ever - because nowadays people can copy & paste others' work without attribution more readily, and untrained listeners have no idea of the ripoff; such as, cloning an entire iGEM project as a VC pitch.


This is from the project's site http://clipped.me/athelaslanding/

 1. Launch the Athelas app and begin the microscopy.

2. The attached lens will magnify and focus on the sample.

3. Take the picture - and it's sent to our backend to process, count, and classify the cells. The results are then spit out.

This solution may be one of the most expensive microscopes on the market for it's optics performance since it requires a $600 smart phone to work.  Oh, it's targeted towards developing countries?  You mean, those which don't even have smart phones?

These are a couple quotes from the YC comment page.  Which basically prove that it is easy to fool the YC judges with mock science projects.

If you're asking about the competition I was in, with the "miracle device," I forgot to mention that they gave multiple examples of what it could detect. Sore throat, ear infection, etc. ... this just "magically" knew that you had Streptococcus pneumoniae or Haemophilus influenzae present--no demo, no explanation of how.

Here's what I used: http://www.edmundoptics.com/optics/optical-lenses/ball-conde...

http://www.edmundoptics.com/optics/optical-lenses/ball-condenser-lenses/n-bk7-ball-lenses/2041

  Basically, I took a piece of rubber, poked a hole in it, and fit the lens in. It took a couple of hours to get the positioning right but after that it worked like a charm :)    ...
The iPhone camera flash turned out to work best.

From what I remember it was an iPhone with the flash light turned on face down on the table and then a toilet paper roll on top of it with a blood sample in one of those rectangular glass things on top of the toilet paper roll with another iPhone examining it. Please correct me if I'm wrong

He said he used training data sets online to test.

There are several commercial microscopy lens solutions for iPhones now which are very good and relatively low cost.  (I don't have reference links handy, maybe someone can follow up.)


## Jonathan Cline
## jcl...@ieee.org
## Mobile: +1-805-617-0223
########################

[Will Sutton]

Jonathan, 

You make some really good points. But:

• I actually don't mean "good engineering",that's an unrealistic expectation. 30 hours of work is by necessity.

• People who were there, people who read about it might now say - Oh, that wasn't too hard, I have a similar skill level, maybe I could try something like that, but I know how to overcome problem X. That's the key outcome, awareness, exposure. Not the deliverable.
  

By the way, everyone's favorite glowing plant has been accepted into YC incubator: TechCrunch, HN