4 ms·

I'd agree. 20 years ago comments like "we'll have the evolutionary tree of life (or even small clades) resolved in 5-10 years" were routinely tossed around. Tod

by xipho 6y ago

I'd agree. 20 years ago comments like "we'll have the evolutionary tree of life (or even small clades) resolved in 5-10 years" were routinely tossed around. Today the reality of the vastness of the biodiversity, the complexity and limitations of the genome (homoplasy, etc.), limitations in compute power (it takes weeks to run single evolutionary reconstruction analyses), and the and practical inability to scale to 10s of thousands of organisms (things like basic project-tracking software are required) remain huge bottlenecks. Furthermore those "genomes' we have, even outside a handful of model organisms, are more or less raw data, not annotated, confused jumbles of pipe-line derived data that require years to fully grok, refine etc. I suspect we have decades to go before the collective "genomic" enterprise will be "fully operational".

dnautics 6y ago
hm. I somewhat disagree. We basically have the evolutionary tree of life done. There are probably a few surprises still. 90% of stuff, it turns out, is the same. We don't know what causes detailed differences, (like what is different between a dog and a bear?) but for a lot of interesting tasks: "get a picture of how X gene is regulated", "understand how X molecule is biosynthesized" are basically solved, solvable, or there is a worked out procedure to solve it, using sequencing... The problem is that we don't have operators smart enough to know how to use these data carefully. If you are a detail-oriented biochemist who has a good grounding in first principles, you basically have everything you could need out of genomics. However, that won't be enough to get a faculty position!
- xipho 6y ago
  As a point of reference I'm currently a Co-PI on a grant to recover the phylogeny of a hyper-diverse group of insects. As to your first point- we are nowhere close to a complete evolutionary tree of life (assuming we could come up with what "complete" means). I agree we have good estimations of some basic major splits, but many things are in flux. Early origins? "Kingdom" classifications change multiple times each year still. Your second point (aside from "people are dumb", which doesn't really hold) is not about evolution, and I'm not experienced enough to make a judgement. I strongly doubt, however, that the full capacity of "compute" you need to actually do genomics, is anywhere near enough to answer everything we'd want to. Today you need a full team of support, not just one person with grounded first principles (which I agree is often missing, but it's also required of more than one expert type).
  - dnautics 6y ago
    I guess if you're an evolutionary biologist the thing is that there will always be more to discover from a genomics perspective (so it's a bit of a search for smaller and smaller), but if you step back and do quantitative analysis of the structure of the tree of life I think it will be the case that what we discover in the next 100 years is not going to be as dramatic as what we have learned in the last 100. Big shifts (like archae; my grad school boss discovered that methanogens are archae) came before genomics even! And while they are still some surprises, they are increasingly on the edges. For many things you probably need to only do a blast search. for example, I identified candidate mutations in an enzyme one combination of which dramatically improved the output by carefully doing a blast search. Likewise, my postdoc boss basically figured out an entire small molecule biosynthetic pathway with ~ a day's worth of blast searching. To understand transcriptional networks, you can get really far with qpcr, or if you want to get really old school yeast-two-hybrid, both of these impossible without genomics. We have total organism knockout Libraries in yeast and e coli that are powerful tools, and have a genome that we've minimized to sub-500 genes. Yes there's 100 or so that we don't understand (mostly short transmembrane proteins that honestly probably need to be there in bulk just to maintain tonicity). These are just the projects I've directly worked on or been a fly-on-the-wall on. Anyways, point is most of the really bid things that the genomics revolutionized have been ticked off. There's still a lot to do, obviously, but I think we know more than you're giving credit for. It does nobody a service to treat biology as a mysterious black box that will never be penetrated.

by xipho 6y ago

dnautics 6y ago

hm. I somewhat disagree. We basically have the evolutionary tree of life done. There are probably a few surprises still. 90% of stuff, it turns out, is the same. We don't know what causes detailed differences, (like what is different between a dog and a bear?) but for a lot of interesting tasks: "get a picture of how X gene is regulated", "understand how X molecule is biosynthesized" are basically solved, solvable, or there is a worked out procedure to solve it, using sequencing... The problem is that we don't have operators smart enough to know how to use these data carefully. If you are a detail-oriented biochemist who has a good grounding in first principles, you basically have everything you could need out of genomics. However, that won't be enough to get a faculty position!

xipho 6y ago
As a point of reference I'm currently a Co-PI on a grant to recover the phylogeny of a hyper-diverse group of insects. As to your first point- we are nowhere close to a complete evolutionary tree of life (assuming we could come up with what "complete" means). I agree we have good estimations of some basic major splits, but many things are in flux. Early origins? "Kingdom" classifications change multiple times each year still. Your second point (aside from "people are dumb", which doesn't really hold) is not about evolution, and I'm not experienced enough to make a judgement. I strongly doubt, however, that the full capacity of "compute" you need to actually do genomics, is anywhere near enough to answer everything we'd want to. Today you need a full team of support, not just one person with grounded first principles (which I agree is often missing, but it's also required of more than one expert type).
- dnautics 6y ago
  I guess if you're an evolutionary biologist the thing is that there will always be more to discover from a genomics perspective (so it's a bit of a search for smaller and smaller), but if you step back and do quantitative analysis of the structure of the tree of life I think it will be the case that what we discover in the next 100 years is not going to be as dramatic as what we have learned in the last 100. Big shifts (like archae; my grad school boss discovered that methanogens are archae) came before genomics even! And while they are still some surprises, they are increasingly on the edges. For many things you probably need to only do a blast search. for example, I identified candidate mutations in an enzyme one combination of which dramatically improved the output by carefully doing a blast search. Likewise, my postdoc boss basically figured out an entire small molecule biosynthetic pathway with ~ a day's worth of blast searching. To understand transcriptional networks, you can get really far with qpcr, or if you want to get really old school yeast-two-hybrid, both of these impossible without genomics. We have total organism knockout Libraries in yeast and e coli that are powerful tools, and have a genome that we've minimized to sub-500 genes. Yes there's 100 or so that we don't understand (mostly short transmembrane proteins that honestly probably need to be there in bulk just to maintain tonicity). These are just the projects I've directly worked on or been a fly-on-the-wall on. Anyways, point is most of the really bid things that the genomics revolutionized have been ticked off. There's still a lot to do, obviously, but I think we know more than you're giving credit for. It does nobody a service to treat biology as a mysterious black box that will never be penetrated.