The Team, The Team, The Team: Reductionism vs holism in microbiome research (CC274)

This is a seminar that Pat gave in his home department at the University of Michigan. Even though it isn't a data science or reproducible research talk, I think you'll find the tension between reductionism and holism interesting. Pat is definitely on the holism end of the spectrum.
You can find my blog post for this episode at www.riffomonas...

[@r4lpha]

Thanks for sharing Pat. 100% agree with the point about people wanting a 'Silver Bullet'. Note that so called 'high-impact' journals (e.g. CNS) tend to promote this also, as it makes easier headlines.
As a microbial ecologist I think that you have to accept uncertainty and complexity -- something that people from more traditional scientific backgrounds are uncomfortable with.

[@Riffomonas]

Absolutely - one of my favorite "jokes" divides those who can deal with complexity and those who can't. What do you call two strains of bacteria living together? Ecology! Contamination! Take your pick 😂

[@Lia-Turner]

I love the idea of a gestalt approach to studying Biology. Wonderful talk. Thank you for sharing.

[@Riffomonas]

Thanks! I like how you put it by saying "gestalt approach" - that's great

[@man_shooo]

Thanks a lot for following up on this! Great talk. I knew it the moment I saw the title that it would be related to the "biologist fix a radio" discussion.
Two points:
- It makes sense why there is such an exponential increase in the number of papers using ML methods in biology. Some of it might indeed be hype, but these methods really do provide a multi-variable/holistic look at complex biological systems.
- Re biological researchers' disbelief in stats and quantitative methods -- I have also experienced this first-hand during my undergrad. It is a bit ironic, given the first point I mentioned, but I think there is still a lot to gain/learn about such systems if we simply embrace and train people more quantitatively.

[@Riffomonas]

ha! you found it :) Thanks for watching and sharing your thoughts.

[@eric13hill]

Too much of microbiology focuses on naming individual organisms as though doing so is a sufficient end or purpose. Even worse is collecting groups of names as though that is useful. I'd rather see collections of functions, unique genetic abilities, or their absence rather than organism names. To do something useful with this information, tell us what functions are collectively present with this group of organisms that are not present in this other group. We need to make the transition toward higher level conclusions from the data. What are the microbes doing, or not doing? How are they doing what they do?

[@Riffomonas]

Thanks for your comment! One danger I have seen in lists of functions (aka metagenomics) is the assumption that all functions are active wherever their genes are found. For that we need to get to transcriptomics and eventually proteomics.

[@eric13hill]

@@Riffomonas Epigenetics can give some insights on products of genes being different than expected. i think looking for commonalities in function will reveal more of the mechanisms masked by complexity.

[@drjzee]

Not sure where I fit on reductionism vs. holism, except that I think that both have their merits. I have spent time in both and been affected by the 'descriptive' response in reviews. So much so that I now think it rises to a rather lazy ad-hominem critique sometimes, which is certainly irritating in a grant review. This kind of comment on an application also misses the important point that how we can understand the world without describing it? It also misses that many important clinical advances have come out of understanding diseases by describing them.

[@Riffomonas]

Thanks for watching and sharing!