4 ms·

Interesting. What would be one of those factual statements given your expertise in the area that the ML folks don't understand?

by geertj 4y ago

dekhn 4y ago
"Improved protein structure prediction will speed up drug discovery"
- melling 4y ago
  Explaining why this won’t happen is the obvious follow up question. When is “improved structure prediction” useful or important? Often a process is simplified or distilled down to a sound bite for the general population. Then we simply repeat without understanding the details.
  - dekhn 4y ago
    Improved structure prediction is mainly useful in hypothesis generation when doing hypothesis-driven science (IE, you want to confirm that a specific part of a protein plays a functional role in a disease). Its also a nice way to think creatively about your protein of interest. THe problem is those distilled soundbites get learned by the next generation and they try to apply it. At least I will give AlphaFold/DM credit for correcting their language - originally they claimed AF solved protein folding, but really, it's just a structure predictor, which is an entirely different area. Unfortunately, people basically taught computer scientists that the Anfinsen Dogma was truth. I fell for this for many years.
    - warinukraine 4y ago
      https://en.wikipedia.org/wiki/Anfinsen%27s_dogma https://en.wikipedia.org/wiki/Anfinsen%27s_dogma > It states that, at least for a small globular protein in its standard physiological environment, the native structure is determined only by the protein's amino acid sequence. Seems like "no true scotsman". If you present a counter example, they'll go "but this is only true for "small", the one you gave me isn't small.
      - dekhn 4y ago
        Let's say you have a pool of smart first year grad students you want to inspire to work hard on problems for you for the next 7 years. Do you say "we're going to give you a problem that is unsolved but likely has a general solution, and you have a chance of making progress, publishing, and moving on to a postdoc" or do you say "THis is an impossibly hard problem and you will only make a marginal improvement on the state of the art because the problem space is so complex and large"? You say the first because it gets the students interested and working on the problem, only to learn many years later that the simplified model presented was so simplified it wasn't helpful. I fell for that and spent years working on drug discovery, structure prediction, etc, only to realize: while what Anfinsen said was true, it only applies to about 1% of protein space. It's not so much a "no true scotsman" as "some scotsmen wear quilts, and others have beards, but neither of those is sufficient to classify an example as scots".
- alextheparrot 4y ago
  The core caveat being the yield of “improved … prediction” in the future tense, I assume? Given that improved structures has sped up drug discovery, I can see where the mistake is made (X has improved Y, therefore X will improve Y)
- ProjectArcturis 4y ago
  It won't? AI-powered drug screening has definitely been overhyped, but in the longer term highly accurate protein structure modelling should let us understand protein-protein interactions and provide new opportunities for intervention.
  - dekhn 4y ago
    That has been a long-stated claim in the field that has repeatedly not shown to be useful in any sort of "engineering" or "medical treatment" sense.
    - gameman144 4y ago
      Hrmm, I think I was expecting a different type of disagreement when you said he was wrong on facts. I'm sure you know your stuff, and that you have a lot of experience with proteins that haven't helped with drug discovery or engineering, but it sounds like this is indeed a mismatch between predictions rather than facts. It very well could be the case that speeding up certain problems by multiple orders of magnitude really does help with drug discovery, and this isn't factually inconsistent with the fact that solving those problems hasn't turned out to be useful so far in this area.
      - lesuorac 4y ago
        If you can discover a drug faster and that drug is as useful as dirt does it matter? This isn't my field but I could grab a bunch of random jars off a shelf and pour them into capsules. No matter how fast I can do this won't improve medical outcomes for patients.
        dekhn 4y ago
        you just described how modern high content screening, which has been one of the most useful techniques for finding drug leads, works. Since lead-finding is a bottleneck in the drug discovery process, it has been highly effective because it can measure things that are not currently computationally accessible.

Interesting. What would be one of those factual statements given your expertise in the area that the ML folks don't understand?

by geertj 4y ago

Interesting. What would be one of those factual statements given your expertise in the area that the ML folks don't understand?

dekhn 4y ago

"Improved protein structure prediction will speed up drug discovery"

melling 4y ago
Explaining why this won’t happen is the obvious follow up question. When is “improved structure prediction” useful or important? Often a process is simplified or distilled down to a sound bite for the general population. Then we simply repeat without understanding the details.
- dekhn 4y ago
  Improved structure prediction is mainly useful in hypothesis generation when doing hypothesis-driven science (IE, you want to confirm that a specific part of a protein plays a functional role in a disease). Its also a nice way to think creatively about your protein of interest. THe problem is those distilled soundbites get learned by the next generation and they try to apply it. At least I will give AlphaFold/DM credit for correcting their language - originally they claimed AF solved protein folding, but really, it's just a structure predictor, which is an entirely different area. Unfortunately, people basically taught computer scientists that the Anfinsen Dogma was truth. I fell for this for many years.
  - warinukraine 4y ago
    https://en.wikipedia.org/wiki/Anfinsen%27s_dogma https://en.wikipedia.org/wiki/Anfinsen%27s_dogma > It states that, at least for a small globular protein in its standard physiological environment, the native structure is determined only by the protein's amino acid sequence. Seems like "no true scotsman". If you present a counter example, they'll go "but this is only true for "small", the one you gave me isn't small.
    - dekhn 4y ago
      Let's say you have a pool of smart first year grad students you want to inspire to work hard on problems for you for the next 7 years. Do you say "we're going to give you a problem that is unsolved but likely has a general solution, and you have a chance of making progress, publishing, and moving on to a postdoc" or do you say "THis is an impossibly hard problem and you will only make a marginal improvement on the state of the art because the problem space is so complex and large"? You say the first because it gets the students interested and working on the problem, only to learn many years later that the simplified model presented was so simplified it wasn't helpful. I fell for that and spent years working on drug discovery, structure prediction, etc, only to realize: while what Anfinsen said was true, it only applies to about 1% of protein space. It's not so much a "no true scotsman" as "some scotsmen wear quilts, and others have beards, but neither of those is sufficient to classify an example as scots".
alextheparrot 4y ago
The core caveat being the yield of “improved … prediction” in the future tense, I assume? Given that improved structures has sped up drug discovery, I can see where the mistake is made (X has improved Y, therefore X will improve Y)
ProjectArcturis 4y ago
It won't? AI-powered drug screening has definitely been overhyped, but in the longer term highly accurate protein structure modelling should let us understand protein-protein interactions and provide new opportunities for intervention.
- dekhn 4y ago
  That has been a long-stated claim in the field that has repeatedly not shown to be useful in any sort of "engineering" or "medical treatment" sense.
  - gameman144 4y ago
    Hrmm, I think I was expecting a different type of disagreement when you said he was wrong on facts. I'm sure you know your stuff, and that you have a lot of experience with proteins that haven't helped with drug discovery or engineering, but it sounds like this is indeed a mismatch between predictions rather than facts. It very well could be the case that speeding up certain problems by multiple orders of magnitude really does help with drug discovery, and this isn't factually inconsistent with the fact that solving those problems hasn't turned out to be useful so far in this area.
    - lesuorac 4y ago
      If you can discover a drug faster and that drug is as useful as dirt does it matter? This isn't my field but I could grab a bunch of random jars off a shelf and pour them into capsules. No matter how fast I can do this won't improve medical outcomes for patients.
      - dekhn 4y ago
        you just described how modern high content screening, which has been one of the most useful techniques for finding drug leads, works. Since lead-finding is a bottleneck in the drug discovery process, it has been highly effective because it can measure things that are not currently computationally accessible.