6 ms·

Great! This is why I went back to school to get my biochem degree. I was a data scientist working with hadoop + mahout at a startup and I said "this is going to

by pvnick 12y ago

Great! This is why I went back to school to get my biochem degree. I was a data scientist working with hadoop + mahout at a startup and I said "this is going to revolutionize medicine." When I get my masters in biomedical engineering in a few years I have a dream to help found a company that applies big data and machine learning to problems with clinical significance. Elizabeth, you may end up hearing from me :P

lanaius 12y ago
If startups don't work out because of sensitivity in medical data, look at Northrop Grumman. They work with the CDC on big data analysis.
dbcooper 12y ago
Look for something beyond target identification, if you can. That is lowest part of the value chain.
- pvnick 12y ago
  Not sure what you mean by targets, are you talking about enzymes and drug discovery?
  - dbcooper 12y ago
    Typically I would be referring to the protein (or other molecule) that you want to target with a drug, to either reduce or promote its action. As wiki puts it: http://en.wikipedia.org/wiki/Drug_discovery#Drug_targets http://en.wikipedia.org/wiki/Drug_discovery#Drug_targets Generally, the "target" is the naturally existing cellular or molecular structure involved in the pathology of interest that the drug-in-development is meant to act on.
- antifragile 12y ago
  True, 'cause one drug for one condition now costs $4-12B, because the success rate is < 0.02%. There needs to be multi-functionality in our innovative efforts when we experiencing such high rates of disruption, say, extinctions.
danieltillett 12y ago
This is an amazingly hard problem and I wish you all the best. One thing you won't learn at school is all the complications involved in validating a target. The best resource I know of for this is the "in the pipeline" blog by Derek Lowe. I have spent 100s of hours reading all his posts (they go back 10 years) and especially the comments. It is a hard slog, but the amount you will learn is staggering. Edit. I guess is should include the link http://pipeline.corante.com/ http://pipeline.corante.com/
- antifragile 12y ago
  I feel an interesting and charged buzz around this thread, so though to join in. Things appear hard and severely constrained because we habitually fall into the trap of thinking linearly and incrementally, as opposed to, at least sometimes, multi-linearly and in bounds, by perspective shifting. To do that one has to question the fundamental assumptions of which this linear trajectory were founded on. More often than not, one finds that much of what is considered so is merely based on someone', usually a group's very convincing, albeit not necessarily correct narrative.
dnautics 12y ago
good luck. Don't forget Sydney Brenner's warning about high-throughput (could also be said of bioinformatics): [it's] low-input, high-throughput, no-output biology
wuschel 12y ago
Combining the scalibility of software startups with biotechnology is a very interesting approach, indeed. Apart from the different timeframe when it comes to working with software and wetware (e.g. lab work with biomaterials), I wonder how easy it will be to get people and know-how on board. In Europe, young scientists (speak: PhD-slaves) are often bound to their research institutes when it comes to the exploitation of the IP generated by their research. They may end up with publications or patents, but usually it is the partners from industry that bring the product to the market. How does this compare to the U.S.?
antifragile 12y ago
Such vision. Good luck to you. Otherwise, maybe I'll have a cool job for you. :-)

by pvnick 12y ago

lanaius 12y ago

If startups don't work out because of sensitivity in medical data, look at Northrop Grumman. They work with the CDC on big data analysis.

dbcooper 12y ago

Look for something beyond target identification, if you can. That is lowest part of the value chain.

pvnick 12y ago
Not sure what you mean by targets, are you talking about enzymes and drug discovery?
- dbcooper 12y ago
  Typically I would be referring to the protein (or other molecule) that you want to target with a drug, to either reduce or promote its action. As wiki puts it: http://en.wikipedia.org/wiki/Drug_discovery#Drug_targets http://en.wikipedia.org/wiki/Drug_discovery#Drug_targets Generally, the "target" is the naturally existing cellular or molecular structure involved in the pathology of interest that the drug-in-development is meant to act on.
antifragile 12y ago
True, 'cause one drug for one condition now costs $4-12B, because the success rate is < 0.02%. There needs to be multi-functionality in our innovative efforts when we experiencing such high rates of disruption, say, extinctions.

danieltillett 12y ago

This is an amazingly hard problem and I wish you all the best. One thing you won't learn at school is all the complications involved in validating a target. The best resource I know of for this is the "in the pipeline" blog by Derek Lowe. I have spent 100s of hours reading all his posts (they go back 10 years) and especially the comments. It is a hard slog, but the amount you will learn is staggering. Edit. I guess is should include the link http://pipeline.corante.com/ http://pipeline.corante.com/

antifragile 12y ago
I feel an interesting and charged buzz around this thread, so though to join in. Things appear hard and severely constrained because we habitually fall into the trap of thinking linearly and incrementally, as opposed to, at least sometimes, multi-linearly and in bounds, by perspective shifting. To do that one has to question the fundamental assumptions of which this linear trajectory were founded on. More often than not, one finds that much of what is considered so is merely based on someone', usually a group's very convincing, albeit not necessarily correct narrative.

dnautics 12y ago

good luck. Don't forget Sydney Brenner's warning about high-throughput (could also be said of bioinformatics): [it's] low-input, high-throughput, no-output biology

wuschel 12y ago

Combining the scalibility of software startups with biotechnology is a very interesting approach, indeed. Apart from the different timeframe when it comes to working with software and wetware (e.g. lab work with biomaterials), I wonder how easy it will be to get people and know-how on board. In Europe, young scientists (speak: PhD-slaves) are often bound to their research institutes when it comes to the exploitation of the IP generated by their research. They may end up with publications or patents, but usually it is the partners from industry that bring the product to the market. How does this compare to the U.S.?

antifragile 12y ago

Such vision. Good luck to you. Otherwise, maybe I'll have a cool job for you. :-)