At first glance, the list of animals could suggest any zoo. There’s an elephant, an armadillo, an opossum, a dolphin, a sloth, a hedgehog, big and small bats, a couple of shrews, some fish, a macaque, an orangutan, a chimpanzee and a gorilla—to name a few of the more familiar creatures. But this menagerie is not at all like any zoo that has been constructed before. There are no cages, no concession stands and, in fact, no animals. It is a “virtual” zoo that contains only the DNA sequences of those animals—the hundreds of millions to billions of letters of DNA code that make up the genetic recipe for each species.

The most excited visitors to this new molecular zoo are evolutionary biologists, because within it lies a massive and detailed record of evolution. For many decades, scientists have longed to understand how the great diversity of species has arisen. We have known for half a century that changes in physical traits, from body color to brain size, stem from changes in DNA. Determining precisely what changes to the vast expanse of DNA sequences are responsible for giving animals their unique appearance was out of reach until recently, however.

This article was originally published with the title Regulating Evolution.

7 Comments

1. 1. SteveBallmer 01:18 PM 5/5/08

   I don't know if I can swallow all of that

   Reply | Report Abuse | Link to this

2. 2. sampablokuper 04:18 PM 5/5/08

   "It is a virtual zoo..." Where is it? Are the DNA sequences publicly accessible? Please can you provide a hyperlink?

   Reply | Report Abuse | Link to this

3. 3. frgough 12:47 PM 5/6/08

   It seems that this finding actually bolster's Behe's arguments on irreducible complexity. Before you can get effective evolution, you have to have a gene and helper gene structure in place, and those have to occur by pure random chance before selection can start working to bias the probabilities.
   
   I would suspect a mathematical analysis of this model would make Haldane's number even higher.

   Reply | Report Abuse | Link to this

4. 4. I'm no "cis"sy 09:49 PM 5/6/08

   There are a number of evolutionary biologists that remain unconvinced that MOST evolution proceeds through changes in on/off switches for genes, also known as cis-regulatory evolution (see Hoekstra and Coyne, 2007 in the journal Evolution). Carroll and company fail to address the concerns of this community. Here are a few reasons, not mentioned in this article, to be skeptical of Carroll et al's claims:
   1. Genes often duplicate: this means that one copy of a gene can sustain all the old, necessary interactions and the new copy can evolve new functions.
   2. Scientists have not yet identified the exact mutations that explain an evolved difference in gene regulation (although many of the examples in this paper appear to be caused by gene regulation, we can't be sure until they find a causal DNA sequence).
   3. Protein structure = biochemistry. To evolve a novel chemical reaction you need to change coding-DNA...this is a major trend of organic evolution that Carroll et al underestimate.

   Reply | Report Abuse | Link to this

5. 5. SeanBCarroll 04:24 PM 5/30/08

   Here are some replies to comments posted:
   
   1)To sampablokuper
   ""It is a virtual zoo..." Where is it? Are the DNA sequences publicly accessible? Please can you provide a hyperlink?"
   
   Yes, the sequences are publicly accessible.
   Go to http://www.ensembl.org/index.html
   
   To I'm no "cis"sy
   2)There are a number of mistaken impressions and statements in this comment.
   "There are a number of evolutionary biologists that remain unconvinced that MOST evolution proceeds through changes in on/off switches for genes, also known as cis-regulatory evolution (see Hoekstra and Coyne, 2007 in the journal Evolution). Carroll and company fail to address the concerns of this community."
   
   I have never stated nor do I think that MOST evolution occurs via cis-regulatory sequences. This artcile, most of my laboratory's work, and previous reviews I have written have focused on MORPHOLOGICAL EVOLUTION which, because it involves genes with special properties such as many enhancers and many independent functions, is more likely to proceed through cis-regulatory sequences. Hoekstra and Coyne blurred the distinction between morpholoigical evolution in particular - my focus - and adaptation and evolution in general. This was unfortunate. Furthermore, I should note that I have written an entire book chronicling examples of organismal evolution that involved gene duplication and/or the evolution of coding sequences ( The Making of the Fittest, 2006)
   
   ii) Here are a few reasons, not mentioned in this article, to be skeptical of Carroll et al's claims:
   1. Genes often duplicate: this means that one copy of a gene can sustain all the old, necessary interactions and the new copy can evolve new functions."
   
   Yes, genes do get duplicated. But ,as I have pointed out before, these events are infrequent in morphology-regulating genes relative to the long span of animal evolution. For example, no Hox genes have arisen in the past 400 million years of tetrapod evolution that gave us frogs, birds, snakes, turtles, giraffes, etc.
   
   ii. Scientists have not yet identified the exact mutations that explain an evolved difference in gene regulation (although many of the examples in this paper appear to be caused by gene regulation, we can't be sure until they find a causal DNA sequence).
   
   This statement is incorrect. Evolutionary changes in binding sites for key transcription factors have been documented in, for example:
   Gompel et al. 2005 Nature 433:481-487
   Jeong et al 2006 Cell 125:1387-1399
   Tournamille et al 1995 Nature Genetics 10 : 224-228
   
   iii) Protein structure = biochemistry. To evolve a novel chemical reaction you need to change coding-DNA...this is a major trend of organic evolution that Carroll et al underestimate."
   
   Both parts of this statement are also mistaken. Protein-DNA interactions can evolve by changing the protein or the target DNA sequence. The former is rare, the latter is pervasive. See Tuch et al. 2008 PLoS Biology Feb 6(2): e38 and references therein to prior work by the Johnson lab.
   
   Regarding the assertion that I have underestimated protein evolution, this is contradicted by my extensive discussion of all sorts of examples involving opsins, antifreezes, globins, ribonuclease, pigmentation proteins etc
   
   The failure here, as with Hoeksra and Coyne, is to make the critical disticntion between morphological evolution and evolution in general when assessing what types of molecular changes are most meaningful and prevalent.

   Reply | Report Abuse | Link to this

6. 6. FredC 09:36 AM 8/8/08

   But isn't color an aspect of morphology, too? You are very quick to adduce color and pattern differences as evidence for your theory when they are due to a cis-regulatory mutation (as in the difference between the two species of Drosophila), but not when the color differences are due to a structural mutation (as in mice). Please clarify whether color is or is not a "morphological" trait. Thank you.

   Reply | Report Abuse | Link to this

7. 7. cbutleruf in reply to SeanBCarroll 06:51 PM 8/10/10

   Are the switches added to live flies or to the genetics of developing ones? If they are being added to live ones, can results occur by adding switches to different but similar species? (and what about other switches between othere species?)

   Reply | Report Abuse | Link to this