Like a subway map, for SNIPs
March 25, 2005 4:26 AM Subscribe
Pretty and pretty interesting: unrooted haplotype networks -- diagrams showing the relation and mutational distance between different sets of DNA, with haplotypes represented by circles proportional to haplotype frequency, joined by lines proportional to mutational difference between haplotypes -- in cichlid fish (on page 3 ) [pdf], in stone loach fish ( on page 3) [pdf], in lesser prairie chickens (on page 6) [pdf] and in a ring species! (on page 2) [pdf]
At the risk of offending you orthogonality, which would certainly not be my intention, can I just make the constructive observation that the text of your post, even to a Sci. grad such as myself, is somewwhat obtuse and not self-revealing in any major way and I wonder if other travellers through our site wouldn't be more interested and sufficiently attracted to examine the links if the preamble was framed in a more generally descriptive manner with less of the esoteric Mol.bio language. I donot want to make this a MeTa topic and as I say, it is meant to be a constructive aside, but it occurs to me now, especially following a somewhat similar style of FPP with the Pruitt cress posting from the other day, which I very much enjoyed. And of course, I realize people are free to click or not click whichever the hell links appeal. I will be examining these in some detail anyway.
It is merely a suggestion for future consideration. :-)
posted by peacay at 9:25 AM on March 25, 2005
It is merely a suggestion for future consideration. :-)
posted by peacay at 9:25 AM on March 25, 2005
Fascinating charts, but the ring species examples have been disproven.
The Herring Gulls: The ring is more than one species with complex patterns of reproductive isolation and hybridization.
Irwin, D. E., and Irwin, J. H. 2002. Circular overlaps: rare demonstrations of speciation. The Auk 199(3):596-602.
The salamanders: The ring is actually 11 very distinct species.
Coyne, J. A., and Orr, H. A. 2004. Speciation. Sinauer Associates, Sunderland, MA.
The greenish warblers: The populations east and west of Kashmir are probably two species that diverged in allopatry.
Coyne, J. A., and Orr, H. A. 2004. Speciation. Sinauer Associates, Sunderland, MA.
posted by estelahe at 10:42 AM on March 25, 2005
The Herring Gulls: The ring is more than one species with complex patterns of reproductive isolation and hybridization.
Irwin, D. E., and Irwin, J. H. 2002. Circular overlaps: rare demonstrations of speciation. The Auk 199(3):596-602.
The salamanders: The ring is actually 11 very distinct species.
Coyne, J. A., and Orr, H. A. 2004. Speciation. Sinauer Associates, Sunderland, MA.
The greenish warblers: The populations east and west of Kashmir are probably two species that diverged in allopatry.
Coyne, J. A., and Orr, H. A. 2004. Speciation. Sinauer Associates, Sunderland, MA.
posted by estelahe at 10:42 AM on March 25, 2005
peacay writes "At the risk of offending you orthogonality, "
Put very gently, and you're absolutely right.
My problem was, I wanted to avoid "Get Your Own Blog Fuckwit" call-outs, so I didn't want to write my own copy describing what an unrooted haplotype network is. I spent quite literally several hours looking for a page with a decent definitions of what an unrooted haplotype network is, and couldn't find anything (that wasn't subscriber only). So in the end, I did write my own description, but tried to keep it as spare as possible.
But as you are too polite to explicitly state, that leaves everyone looking at this and saying <voice="Jon Stewart">"wuh-huh"<Voice="Jon Stewart">?
I was kinda hoping a biologist would happen by and fill in the lacunae I left, but <voice="Jon Stewart">um, not so good<Voice="Jon Stewart">.
Okay, so: a haplotype is, to a fair approximation, a gene. That is, it's a stretch of DNA, a sequence of (any one of the four) DNA letters, on one of the two chromosomes. (A "gene" would actually be something that produced a protein, while a haplotype is just a sequence, possibly exons , or introns, or both.)
The haplotype network then, shows the variants same stretch of DNA, as taken from samples of individual creatures in the same species. Each of the circles is one haplotype variant. Bigger circles mean more of the individuals had that sequence as opposed to a different sequence. So far, so boring.
What's cool is the lines connecting the haplotype variants. In the networks, the length of the connecting line is proportional to the number of mutations required to change one haplotype to another. In several of the networks linked, each possible variation is shown, either as a filled circle, meaning that that variation was seen in the sampled individuals, or as a dot, meaning that that possible variant wasn't seen.
So say we have three haplotypes, X Y and Z, each with just four DNA base pairs.
X: C C C C
Y: C C C A
Z: C C A A
Then to "get" from X to Y, the haplotype has to change one base pair, the fourth, from a C to an A. So in the diagram, haplotype X would be directly connected to haplotype Y. To "get" from Y to Z would also require a single change, at the third base pair, from C to A. So Z would be directly connected to Y. One way to "get" from X to Z would be to have two mutations, probably in sequence, not at the same time. The first would "move" you yo Y, the second to Z.
That's what the diagrams show: how a "gene" (haplotype) can change, how "near" it is to other genes in the population, and how many members of the population have that gene.
Of especial interest are the little black dots, signifying that to "get" from one place to another, the gene would probably (in the absence of multiple consecutive mutations) have to "be" at the dot's position. But thee are no individuals at the dot position -- presumably because it's not healthy to have that configuration of genes for too very long.
Finally, the networks are called unrooted because they make no assumptions or findings about which haplotype is the ancestor and which the descendent -- the "direction" of "movement" isn't given.
So the diagram shows all the viable sequences, all the ways it's possible to "be", say, a cichlid, and how one variation of "being" a cichlid can give way to a slightly different variant. Every position on the diagram is a way of "being" a cichlid; the big circles are popular way, the small circles are less popular (or healthy) and the black dots are places you (maybe) have to "go" but don't want to "stay" if you want to stay alive.
So it's a sort of "subway map" or diagram for genes. The genes far outlive the cichlids they "travel" in, as they are passed to the cichlid's progeny. But the genes have a history too. We'd see the history in a rooted haplotype diagram -- in the unrooted diagram, we just see the possible "stations" (haplotypes) on the "subway".
The map, then, is the alternatives available to a gene, the paths it can take through what Dawkins calls (after Borges's "Library of Bale" which contained every possible "book" of every sequence of an alphabet of 25 letters, whether the sequence made sense or not) the "Library of Mendel", the mathematical combinatorial of all possible gene sequences. The haplotype network shows which "rooms" in the "Library of Mendel" are inhabitable for this particular stretch of DNA-- and which are not -- or at least that subset which is the rooms in which if you're there, you're a cichlid.
And that's what's fascinating: you look at the diagram and imagine you're gene (haplotype) and you see where you could go (and "be") with a little self-modification of your base pairs.
Oh, and the diagrams in the pdfs are fair more beautiful and full color and contain extra information like where those haplotypes are found geographically, than the ones I could link directly. They're just pretty. And interesting as networks.
posted by orthogonality at 2:05 PM on March 25, 2005
Put very gently, and you're absolutely right.
My problem was, I wanted to avoid "Get Your Own Blog Fuckwit" call-outs, so I didn't want to write my own copy describing what an unrooted haplotype network is. I spent quite literally several hours looking for a page with a decent definitions of what an unrooted haplotype network is, and couldn't find anything (that wasn't subscriber only). So in the end, I did write my own description, but tried to keep it as spare as possible.
But as you are too polite to explicitly state, that leaves everyone looking at this and saying <voice="Jon Stewart">"wuh-huh"<Voice="Jon Stewart">?
I was kinda hoping a biologist would happen by and fill in the lacunae I left, but <voice="Jon Stewart">um, not so good<Voice="Jon Stewart">.
Okay, so: a haplotype is, to a fair approximation, a gene. That is, it's a stretch of DNA, a sequence of (any one of the four) DNA letters, on one of the two chromosomes. (A "gene" would actually be something that produced a protein, while a haplotype is just a sequence, possibly exons , or introns, or both.)
The haplotype network then, shows the variants same stretch of DNA, as taken from samples of individual creatures in the same species. Each of the circles is one haplotype variant. Bigger circles mean more of the individuals had that sequence as opposed to a different sequence. So far, so boring.
What's cool is the lines connecting the haplotype variants. In the networks, the length of the connecting line is proportional to the number of mutations required to change one haplotype to another. In several of the networks linked, each possible variation is shown, either as a filled circle, meaning that that variation was seen in the sampled individuals, or as a dot, meaning that that possible variant wasn't seen.
So say we have three haplotypes, X Y and Z, each with just four DNA base pairs.
X: C C C C
Y: C C C A
Z: C C A A
Then to "get" from X to Y, the haplotype has to change one base pair, the fourth, from a C to an A. So in the diagram, haplotype X would be directly connected to haplotype Y. To "get" from Y to Z would also require a single change, at the third base pair, from C to A. So Z would be directly connected to Y. One way to "get" from X to Z would be to have two mutations, probably in sequence, not at the same time. The first would "move" you yo Y, the second to Z.
That's what the diagrams show: how a "gene" (haplotype) can change, how "near" it is to other genes in the population, and how many members of the population have that gene.
Of especial interest are the little black dots, signifying that to "get" from one place to another, the gene would probably (in the absence of multiple consecutive mutations) have to "be" at the dot's position. But thee are no individuals at the dot position -- presumably because it's not healthy to have that configuration of genes for too very long.
Finally, the networks are called unrooted because they make no assumptions or findings about which haplotype is the ancestor and which the descendent -- the "direction" of "movement" isn't given.
So the diagram shows all the viable sequences, all the ways it's possible to "be", say, a cichlid, and how one variation of "being" a cichlid can give way to a slightly different variant. Every position on the diagram is a way of "being" a cichlid; the big circles are popular way, the small circles are less popular (or healthy) and the black dots are places you (maybe) have to "go" but don't want to "stay" if you want to stay alive.
So it's a sort of "subway map" or diagram for genes. The genes far outlive the cichlids they "travel" in, as they are passed to the cichlid's progeny. But the genes have a history too. We'd see the history in a rooted haplotype diagram -- in the unrooted diagram, we just see the possible "stations" (haplotypes) on the "subway".
The map, then, is the alternatives available to a gene, the paths it can take through what Dawkins calls (after Borges's "Library of Bale" which contained every possible "book" of every sequence of an alphabet of 25 letters, whether the sequence made sense or not) the "Library of Mendel", the mathematical combinatorial of all possible gene sequences. The haplotype network shows which "rooms" in the "Library of Mendel" are inhabitable for this particular stretch of DNA-- and which are not -- or at least that subset which is the rooms in which if you're there, you're a cichlid.
And that's what's fascinating: you look at the diagram and imagine you're gene (haplotype) and you see where you could go (and "be") with a little self-modification of your base pairs.
Oh, and the diagrams in the pdfs are fair more beautiful and full color and contain extra information like where those haplotypes are found geographically, than the ones I could link directly. They're just pretty. And interesting as networks.
posted by orthogonality at 2:05 PM on March 25, 2005
Wow...so this is your alternative FPP?
My mother told me that I would reap more with politeness. ;+ )
But I still haven't had a chance to check out these linx. I've been up all night (yet again) and must punch out some Zzz's now so I shall be back.
You're an I.T. worker aren't you? Why are you getting into the very deep world of genetics ??? - arguably among the most complicated areas of studious endeavour there is - as I've said, I studied Biochem/Mol.Bio and I still have a fair bit of trouble keeping up to the pace, as evidenced with my fumbling comments in the Pruitt post a couple of days ago.
posted by peacay at 2:30 PM on March 25, 2005
My mother told me that I would reap more with politeness. ;+ )
But I still haven't had a chance to check out these linx. I've been up all night (yet again) and must punch out some Zzz's now so I shall be back.
You're an I.T. worker aren't you? Why are you getting into the very deep world of genetics ??? - arguably among the most complicated areas of studious endeavour there is - as I've said, I studied Biochem/Mol.Bio and I still have a fair bit of trouble keeping up to the pace, as evidenced with my fumbling comments in the Pruitt post a couple of days ago.
posted by peacay at 2:30 PM on March 25, 2005
peacay - what's wrong with an IT worker having a background in biological science? Just because you don't make a living from it doesn't mean that it's any less interesting...
(Thus spake the biochemistry graduate who now works in a Life Assurance office :-) )
orthagonality: interesting stuff, although I'm suffering a bit from late-night-information-overload (always a danger of mefi!)...
posted by Chunder at 4:19 PM on March 25, 2005
(Thus spake the biochemistry graduate who now works in a Life Assurance office :-) )
orthagonality: interesting stuff, although I'm suffering a bit from late-night-information-overload (always a danger of mefi!)...
posted by Chunder at 4:19 PM on March 25, 2005
Great stuff, orthog -- I think i'm doing my honours thesis on lake magadi cichlids next year and I haven't read the verheyen yet. Thanks!
posted by docgonzo at 6:35 PM on March 25, 2005
posted by docgonzo at 6:35 PM on March 25, 2005
chunder......absolutely nothing wrong with I.T.-on-Sci-background ........I was just curious is all because I sense it was an interest for orthogonality as opposed to something they studied at University - it is a helluva subject to have an interest in without formally studying it and I'm curious as to why
.............and I worked in insurance for the mostpart since finishing Uni. too.....it just helps slowly dissolve the knowledge away bit by molecular bit
posted by peacay at 9:00 PM on March 25, 2005
.............and I worked in insurance for the mostpart since finishing Uni. too.....it just helps slowly dissolve the knowledge away bit by molecular bit
posted by peacay at 9:00 PM on March 25, 2005
peacay writes "I'm curious as to why"
I'd go into it, but I really don't want (yet another) thread that's all about me. We've had too many of those lately, and that's not what MeFi is about.
(Exceptions of course for those who will rightly point out "this sucks as an FPP", but technically, that can be done while only implying, not explicitly stating "and orthogonality is a dumbass".)
Sorry, I for some reason -- possibly the comp. sci. background peacay alludes to, and comp sci.'s utter absorption with trees, and self-balancing trees, and spanning trees, and trees for this and for that and the next thing, and graphs, and networks -- found the haplotype networks absolutely fascinating and didn't really take into account that, without the background science, they're just colored circles.
posted by orthogonality at 11:25 PM on March 25, 2005
I'd go into it, but I really don't want (yet another) thread that's all about me. We've had too many of those lately, and that's not what MeFi is about.
(Exceptions of course for those who will rightly point out "this sucks as an FPP", but technically, that can be done while only implying, not explicitly stating "and orthogonality is a dumbass".)
Sorry, I for some reason -- possibly the comp. sci. background peacay alludes to, and comp sci.'s utter absorption with trees, and self-balancing trees, and spanning trees, and trees for this and for that and the next thing, and graphs, and networks -- found the haplotype networks absolutely fascinating and didn't really take into account that, without the background science, they're just colored circles.
posted by orthogonality at 11:25 PM on March 25, 2005
last personal comment just to make sure there's no misundersanding.....I was in no way implying 'dumbass' or 'bad FPP' or the like at all....in fact it is the complete opposite, for you're to be commended for tackling such a difficult subject and raising the intellect level around this place. Everything I've written in this thread was meant to be constructive and encouraging and not personally critical. 'nuff sed.
posted by peacay at 11:42 PM on March 25, 2005
posted by peacay at 11:42 PM on March 25, 2005
Oh, after considering your comments further I was reminded of this and note that between this thread and the one on Pruitt, you now know a bunch of Science heads on MeFi to whom you could perhaps bounce ideas or seek clarification in order to construct the type of quality FPP you seem to seek. Collaborative posting is a nifty idea. I'm happy to throw in my 2cents worth via email and this subject matter of course means that an extra day's preparation won't make any difference. I'm sure others following these 2 threads would be equally obliging. Just a thought.
posted by peacay at 12:07 AM on March 26, 2005
posted by peacay at 12:07 AM on March 26, 2005
peacay writes "last personal comment just to make sure there's no misundersanding.....I was in no way implying 'dumbass' or 'bad FPP' or the like at all."
No, no, no! I didn't take what you wrote as criticism at all! It was very encouraging.
I was not saying you said anything like that -- I was admitting someone else could fairly say that, given the lack of clarity of my FPP.
peacay writes "Collaborative posting is a nifty idea. I'm happy to throw in my 2cents worth via email and this subject matter of course means that an extra day's preparation won't make any difference. I'm sure others following these 2 threads would be equally obliging. Just a thought."
It's an excellent idea (as I wrote above, I was hoping for some nice biologist to fill in my lacunae on this one), and one that I'm going to explore in the future. Although I should note I think that all MetaFilter threads have the potential to be collaborative posts, even without pre-arrangement.
Look how much you added to this one, by giving me a chance to clarify what I was trying to do. Look at how much estelahe added about ring species -- had he added links, it would have been a great FPP all by itself (even though my heart will mourn if it turns out ring species do not exist; the concept is just too fascinating).
Ideally, MetaFilter would be not about carping and criticism and call-outs and ideological scalp-taking, but about taking an idea and each from his own expertise or knowledge building it up into an enthralling collaborative journey.
posted by orthogonality at 12:30 AM on March 26, 2005
No, no, no! I didn't take what you wrote as criticism at all! It was very encouraging.
I was not saying you said anything like that -- I was admitting someone else could fairly say that, given the lack of clarity of my FPP.
peacay writes "Collaborative posting is a nifty idea. I'm happy to throw in my 2cents worth via email and this subject matter of course means that an extra day's preparation won't make any difference. I'm sure others following these 2 threads would be equally obliging. Just a thought."
It's an excellent idea (as I wrote above, I was hoping for some nice biologist to fill in my lacunae on this one), and one that I'm going to explore in the future. Although I should note I think that all MetaFilter threads have the potential to be collaborative posts, even without pre-arrangement.
Look how much you added to this one, by giving me a chance to clarify what I was trying to do. Look at how much estelahe added about ring species -- had he added links, it would have been a great FPP all by itself (even though my heart will mourn if it turns out ring species do not exist; the concept is just too fascinating).
Ideally, MetaFilter would be not about carping and criticism and call-outs and ideological scalp-taking, but about taking an idea and each from his own expertise or knowledge building it up into an enthralling collaborative journey.
posted by orthogonality at 12:30 AM on March 26, 2005
Now I've had a good look through the links. And although fascinating, this stuff is very difficult. I'm reminded that, although I loved and did quite well at Mol.Bio, I always found genetics to be incredibly hard. (Some might say that's something of a contradiction but I never got as extensive a grasp on genetics so as to be in a position of sufficient understanding to properly develop the argument)
But the diagrammatic representations in the links are rather cool and for a moment there, I did manage to 'see' what was going on and 'felt' the gene inside the picture, as it were. Rather spooky in an ok way - I used to get such 3rd-eye type feelings about bugs and dna and biochemicals - definitely one of the big plusses of concerted study in the field. [Others think acid is cheaper and requires less work.]
Assimilation of this dense material was 100 times easier with your explication above for which I thank you.
But to be fair to estelahe - they cited material not available to the unpaying netizens which goes to show that it's even more difficult to post a complete FPP on 'deep' science here because it virtually requires some affiliation with a University Library and/or their total access to relevant journals.
Thanks for putting in the hard yards, particularly under such circumstances.
posted by peacay at 6:43 AM on March 26, 2005
But the diagrammatic representations in the links are rather cool and for a moment there, I did manage to 'see' what was going on and 'felt' the gene inside the picture, as it were. Rather spooky in an ok way - I used to get such 3rd-eye type feelings about bugs and dna and biochemicals - definitely one of the big plusses of concerted study in the field. [Others think acid is cheaper and requires less work.]
Assimilation of this dense material was 100 times easier with your explication above for which I thank you.
But to be fair to estelahe - they cited material not available to the unpaying netizens which goes to show that it's even more difficult to post a complete FPP on 'deep' science here because it virtually requires some affiliation with a University Library and/or their total access to relevant journals.
Thanks for putting in the hard yards, particularly under such circumstances.
posted by peacay at 6:43 AM on March 26, 2005
« Older The greatest sexual moments in video game history | 1 + 1 = 2. Really. Honestly. Newer »
This thread has been archived and is closed to new comments
Images linked above: Baltimore/Black-Backed Control Region Unrooted Haplotype Network from Beatrice Kondo's web page, the Finnish human Y chromosome from "Dual Origins of Finns Revealed by Y Chromosome Haplotype Variation", the Middle Eastern human Y chromosome from "The Y Chromosome Pool of Jews as Part of the Genetic Landscape of the Middle East".
posted by orthogonality at 4:28 AM on March 25, 2005