The Panda's Thumb Legacy Archive — pandasthumb.net

Phyloseminar: "Accurate reconstruction of insertion-deletion histories by statistical phylogenetics"

Posted 19 September 2011 by Reed A. Cartwright

Tomorrow at 1:00pm Texas Time (CST), Oscar Westesson will be presenting a Phyloseminar.
The "multiple sequence alignment" is a computational artifact. In nature there is no such thing; rather, an alignment represents a partial summary either of indel history, or of structural similarity. Here we show, via evolutionary simulation tests, that all currently-available multiple alignment tools introduce systematic biases into downstream evolutionary analysis - particularly when used to reconstruct histories of insertions and deletions. I will present our unification of Felsenstein's "pruning" algorithm and "progressive alignment" to build a fast, linearly-scaling approximate-maximum-likelihood phylogenetic alignment/reconstruction algorithm. Inference of evolutionary history in this framework displays a clear improvement in accuracy over non-statistical phylogenetic reconstructions and a massive improvement in performance over slow-running MCMC statistical reconstructions.
Phyloseminars are streamed live, and can be watched anywhere you have an Internet connection. To learn how to connect, visit the Phyloseminar page.

30 Comments

Atheistoclast · 20 September 2011

What is a "pruning algorithm" with respect to phylogenetics? Does it clip the ends off the branches? I have always wondered how you can truly test for phylogenetic reconstructions? How could you falsify such inferential techniques? It does seem to come down to interpretation most of the time.

DS · 20 September 2011

Atheistoclast said: What is a "pruning algorithm" with respect to phylogenetics? Does it clip the ends off the branches? I have always wondered how you can truly test for phylogenetic reconstructions? How could you falsify such inferential techniques? It does seem to come down to interpretation most of the time.
Joe, Here is the reference: JME 17(6):368-378 (1981) Enjoy.

Atheistoclast · 20 September 2011

DS said:
Atheistoclast said: What is a "pruning algorithm" with respect to phylogenetics? Does it clip the ends off the branches? I have always wondered how you can truly test for phylogenetic reconstructions? How could you falsify such inferential techniques? It does seem to come down to interpretation most of the time.
Joe, Here is the reference: JME 17(6):368-378 (1981) Enjoy.
No doubt you read this paper beyond the title.

Reed A. Cartwright · 20 September 2011

If you want to learn about the pruning algorithm, then you should watch the seminar.

DS · 20 September 2011

Atheistoclast said:
DS said:
Atheistoclast said: What is a "pruning algorithm" with respect to phylogenetics? Does it clip the ends off the branches? I have always wondered how you can truly test for phylogenetic reconstructions? How could you falsify such inferential techniques? It does seem to come down to interpretation most of the time.
Joe, Here is the reference: JME 17(6):368-378 (1981) Enjoy.
No doubt you read this paper beyond the title.
No doubt you didn't.

Torbjörn Larsson, OM · 20 September 2011

Wow! I haven't seen biology deniers (or cosmology deniers, or Apollo landing deniers) in a very long time, but on this site they seem to be. I sincerely thought anti-scientists had moved on to deny dark matter or dark energy or physicalism. Well, at least they are disappearing out of the general media scape. d(OvO)b
Atheistoclast said: I have always wondered how you can truly test for phylogenetic reconstructions? How could you falsify such inferential techniques? It does seem to come down to interpretation most of the time.
Let me introduce you to 29+ evidences for evolution. Among other things it provides introduction to how subsets of nested sets (sets of phylogenies) decrease uncertainty in an objectively quantifiable way and allow testing. Incidentally the author, Theobald, published a paper in Nature where he showed that such reconstructions can test the existence of a single cellular LUCA to the likelihood of ~ 1:102040 of error [!]. The existence of a phylogenetic LUCA is many times more likely than the number of atoms in the universe (~ 1080), or as scientists say "a huge number". And this is but one observable biological fact among others. Also, let me point out that all facts and hypotheses are inferred from observations. Testing is a deductive operation (eliminating erroneous facts and theories), and the result of this pruning algorithm [sic!] is observed convergence* so today "The Laws Underlying The Physics of Everyday Life Are Completely Understood". Including chemistry laws (of molecules) and biology laws (of evolution), natch. So you can't tar biology specifically, this is basic to science.

Torbjörn Larsson, OM · 20 September 2011

Oops, C and P error. The missing piece:

----------------
* Which of course can be tested. Going meta, test testing, is both doable and, if you think about it, necessary.

Atheistoclast · 20 September 2011

There is absolutely no chance of a universal common ancestor for all life. Prokaryotic genes, even homologs, are markedly different than eukaryotic ones. Moreover, the two cell types are structurally distinct.

Rumraket · 20 September 2011

Atheistoclast said: There is absolutely no chance of a universal common ancestor for all life. Prokaryotic genes, even homologs, are markedly different than eukaryotic ones. Moreover, the two cell types are structurally distinct.
Will you be supplying the calculation that supports that postulate any time soon or are you just asserting shit into thin air? I'm merely asking.

DS · 20 September 2011

Atheistoclast said: There is absolutely no chance of a universal common ancestor for all life. Prokaryotic genes, even homologs, are markedly different than eukaryotic ones. Moreover, the two cell types are structurally distinct.
Actually, prokaryotic cellular machinery tends to be just simplier versions of the same kinds of molecules and processes that we find in eukaryotes. You know, the ones the creationists claim to be irreducibly complex.

phhht · 20 September 2011

Atheistoclast said: There is absolutely no chance of a universal common ancestor for all life. Prokaryotic genes, even homologs, are markedly different than eukaryotic ones. Moreover, the two cell types are structurally distinct.
One of the four lone arrows in Theistoclast's quiver: denial.

https://www.google.com/accounts/o8/id?id=AItOawm-WhebH0itIDDTj06EQo2vtiF0BBqF10Q · 20 September 2011

IMHO one shouldn't feed atheistoclast. Just google atheistoclast + paper + journal

fnxtr · 20 September 2011

Or, for more LULZ, go here.

Torbjörn Larsson, OM · 21 September 2011

Atheistoclast said: There is absolutely no chance of a universal common ancestor for all life. Prokaryotic genes, even homologs, are markedly different than eukaryotic ones. Moreover, the two cell types are structurally distinct.
Of course they are distinct, since they are different populations. They are also alike, share the same genetic machinery, and that is how biologists conclude that there is no chance of not having a LUCA. You do realize that Nature is peer reviewed, and that Theobald "only" added a valid mathematical method for comparison of a relationship that biologists knew and accepted before his work? Denial of accepted biological science isn't a relevant answer on a science blog. Silence would be more appropriate since you obviously don't know what you are discussing, nor actually interested (which the lack of reference to Theobald's work shows).

Torbjörn Larsson, OM · 21 September 2011

@ A Masked Panda, fnxtr:

Thanks! Obviously I have some catching up to do on the depth of human incompetence. Dunning-Kruger is always an amazing source of oblivious arrogance.

harold · 21 September 2011

Any special reason why three comments responding to the prokaryote/eukaryote thing came through but my extremely similar comment got "moderated"?

harold · 21 September 2011

No paranoia intended :).

RWard · 21 September 2011

I wonder how similar Atheiclast expects lineages to be following 1.5 billion years of divergence?

Rumraket · 21 September 2011

RWard said: I wonder how similar Atheiclast expects lineages to be following 1.5 billion years of divergence?
Who knows? This is the guy you're talking to: http://www.rationalskepticism.org/post1003943.html#p1003937
Atheistoclast said:Your mother played a very limited role other than to act as the vehicle for your creation. The genes from her and your papa provided the information for the production of the proteins that make up 50% of your body and which are necessary for its development and operation. But the direction and organization in the formation of you as a zygote through to an embryo and baby was controlled by a divine agency. Let us throw away our copies of the "Selfish Gene" and open up our Bibles: "For you formed my inward parts; you knitted me together in my mother’s womb." Psalm 139:13 So you see, creationism does not entail believing that species were poofed into existence 6000 years ago but that each and every one of us has been specially created by God. When you realize this, you will stop thinking of Jehovah as a fascist cunt who kills homosexuals and the like, but rather as the One to whom you owe your life and happiness.
Isn't it wonderful? :D

Atheistoclast · 21 September 2011

Torbjörn Larsson, OM said:
Atheistoclast said: There is absolutely no chance of a universal common ancestor for all life. Prokaryotic genes, even homologs, are markedly different than eukaryotic ones. Moreover, the two cell types are structurally distinct.
Of course they are distinct, since they are different populations. They are also alike, share the same genetic machinery, and that is how biologists conclude that there is no chance of not having a LUCA.
I'm pretty sure the aminoacyl synthetases are different - they pair each anticodon on transfer RNA with the corresponding amino acid. Also, the genetic code is different in that prokaryotes have three start sites: ATG,GTG and TTG rather than just one for eukaryotes.

harold · 22 September 2011

I’m pretty sure the aminoacyl synthetases are different - they pair each anticodon on transfer RNA with the corresponding amino acid. Also, the genetic code is different in that prokaryotes have three start sites: ATG,GTG and TTG rather than just one for eukaryotes.
All of this is evidence FOR common descent to any reasonable mind.

DS · 22 September 2011

harold said:
I’m pretty sure the aminoacyl synthetases are different - they pair each anticodon on transfer RNA with the corresponding amino acid. Also, the genetic code is different in that prokaryotes have three start sites: ATG,GTG and TTG rather than just one for eukaryotes.
All of this is evidence FOR common descent to any reasonable mind.
Absolutely. My response to this nonsense is on the bathroom wall. Suffice it to say that phylogenetic analysis, (you know, just one of those things that Joe is completely ignorant of), demonstrates that the ARS enzymes are all related in a nested hierarchy that is consistent with the general shape of the tree of life, (with a few complications, presumably caused by some lateral gene transfer). I can provide references if anyone is genuinely interested. And of course, the canonical genetic code is also strong evidence for a single origin of all extant life forms. Now why am I not surprised that Joe thinks it means just the opposite?

jlesow · 22 September 2011

fnxtr said: Or, for more LULZ, go here.
Well that was a couple hours I'll not get back.

jlesow · 22 September 2011

DS said: Absolutely. My response to this nonsense is on the bathroom wall. Suffice it to say that phylogenetic analysis, (you know, just one of those things that Joe is completely ignorant of), demonstrates that the ARS enzymes are all related in a nested hierarchy that is consistent with the general shape of the tree of life, (with a few complications, presumably caused by some lateral gene transfer). I can provide references if anyone is genuinely interested. And of course, the canonical genetic code is also strong evidence for a single origin of all extant life forms. Now why am I not surprised that Joe thinks it means just the opposite?
Koonin seems to think the amount of HGT is more than just "some." His recent review articles and the book have some interesting arguments. He argues that TOE couldn't address the unicellular tree until genomics was possible. Sequencing seems to be the microscope and telescope for studying evolution.

Joe Felsenstein · 22 September 2011

jlesow said: Koonin seems to think the amount of HGT is more than just "some." His recent review articles and the book have some interesting arguments. He argues that TOE couldn't address the unicellular tree until genomics was possible. Sequencing seems to be the microscope and telescope for studying evolution.
He's sort-of right about that, but one has to stretch the word "genomics" to cover DNA sequencing of the pre-genomic era. Things started to get serious as soon as Carl Woese and Mitch Sogin started using ribosomal RNA sequences, which was about 1978. I'd call that before genomics, but Eugene Koonin might define it differently. Or perhaps he is making the point that one needed many loci (thus almost genome-scale data) to tell HGT from vertical inheritance.

jlesow · 22 September 2011

He gives full credit to Woese, but still thinks that because the genome changes faster than ribosomes, and participates in HGT, it is a better indicator of lineage. (I don't have the background to know if I got that right.)

Torbjörn Larsson, OM · 23 September 2011

Atheistoclast said: I'm pretty sure the aminoacyl synthetases are different -
You are still not responding to the paper. It takes all that into consideration - it has to. This is your 2nd attempt to duck the science. Please stop if you are unable to follow or respond to it.

DS · 23 September 2011

Torbjörn Larsson, OM said:
Atheistoclast said: I'm pretty sure the aminoacyl synthetases are different -
You are still not responding to the paper. It takes all that into consideration - it has to. This is your 2nd attempt to duck the science. Please stop if you are unable to follow or respond to it.
Why? It never stopped him before. All he has to say is that all of authors, reviewers and editors are wrong, only he knows better, end of story. He doesn't even have to give a good reason, just because he says so is good enough.

Rolf · 24 September 2011

harold said:
I’m pretty sure the aminoacyl synthetases are different - they pair each anticodon on transfer RNA with the corresponding amino acid. Also, the genetic code is different in that prokaryotes have three start sites: ATG,GTG and TTG rather than just one for eukaryotes.
All of this is evidence FOR common descent to any reasonable mind.
I am not in a position to say I agree but I feel rather comfortable saying I don't see that as evidence against common descent.

iklimsaklima · 24 October 2011

He gives full credit to Woese, but still thinks that because the genome changes faster than ribosomes, and participates in HGT, it is a better indicator of lineage. (I don’t have the background to know if I got that right.)

klima servisi