Behind the Green Screen
Here’s how we make designer mutations—it takes equipment, hard work, and intelligent intervention. Of course, we could just wait….
Natural Selection's Reach
Mt. Whitney http://flic.kr/p/8Bbuue
A reader wrote us recently to ask why natural selection can’t extract enough information from the fitness landscape to explain complex features. It all depends on what you think the fitness landscape looks like, and what you think the reach of natural selection is. Dr. Douglas Axe explained the nature of the problem in Science and Human Origins, excerpted below:
Darwinian evolution is often thought of in terms of journeys over a vast rugged landscape. Each point on this strange terrain represents a possible genome sequence, those possibilities being so staggeringly numerous that real organisms have only actualized a minute fraction of them. The ground elevation at each point corresponds to the fitness of individuals carrying that genome, with the horizontal distance between any two points indicating the degree to which the corresponding genomes differ. In terms of this picture, all of the millions of species alive today are represented by their own points, high up on peaks scattered somewhere across this conceptual landscape (the fact that they are alive demonstrates the quality of their genomes).
Now, wherever a species happens to be, Darwin’s engine tends to move it toward the highest ground it can reach. According to the Darwinian story, that simple tendency to migrate upward has, over billions of years, transported the first primitive genome from its starting point to higher points along millions of diverging paths. The result is the spectacular variety of life forms we see today with a correspondingly wide dispersal of genomes across the vast conceptual landscape.
But there’s something suspicious about this story. It has to do with the wide disparity of distance scales. The scale of the landscape, which is characterized by the extent to which dissimilar genomes differ, is very large by any reasonable calculation. On the other hand, Darwin’s engine moves in steps that can only reach points a tiny distance away from the prior point. In one step it can move a genome to the highest point within this reach, but further progress would require a still higher point to fall within reach once that move is made.
The problem of climbing in tiny steps. If the rule is move to the highest point that can be reached in each step and the landscape is rugged, then the endpoint will be a local peak. Illustration: Douglas Axe, Science and Human Origins.
That might happen every now and then, but it would have to happen in an amazingly consistent and helpful way to explain how the enormous distances were traversed from the point marking the first primitive organism to the millions of points marking the great variety of modern life forms.
Let’s put this in more familiar terms. The summit of Mount Whitney,the highest point in the contiguous United States, is just 136 kilometers from the lowest point in North America, known as Badwater Basin. Now, suppose there were an automated vehicle capable of remotely scanning the surrounding terrain within some fixed distance and then moving to the highest point identified by the scan. If the scan radius is greater than 136 kilometers, this vehicle could get from Badwater to Whitney in one scan-and-move operation. But what if the scan radius is one millionth that size? Now the circle that the vehicle ‘sees’ from its current position is about a shoe-length across, with each move being up to half that distance. Considering how uneven the ground is, we wouldn’t expect this nearsighted vehicle to complete more than a few scan-and-move operations before becoming stuck on a rock, maybe half a pace from where it started. Summiting Whitney would be completely out of the question. So the idea that any ability to seek higher ground, no matter how restricted, makes the highest summit accessible turns out to be highly simplistic.
Badwater Basin after the rain http://flic.kr/p/7DxHoD
Dr. Axe goes on to explain:
The very same critique applies to Darwinism. Consider that for Darwin’s engine to invent humans from apes, it would have had to work within the severe limitation of a single-mutation scan radius. That is, it would have had to invent humans one simple mutation at a time, with each of these mutations making its possessors significantly more fit than their peers. Contrast this single-mutation reach with the millions of differences that distinguish the chimp and human genomes and we’re back to the impossible trek from Badwater to Whitney. Maybe the genomic landscape is so much simpler and smoother than the Death Valley terrain as to enable Darwin’s engine to cruise upward to exotic destinations on gentle inclines, but why would anyone assume this to be so? Only if experiment after experiment were to prove that remarkable kind of terrain to be the rule should anyone begin to think that something so fantastic might be true.
What is the mutational reach of natural selection in general? It’s very short. For bacteria, our work suggests the reach is only a few mutations at a time. That’s not enough to get a genuinely new function for a protein, let alone a new pathway made up of a handful of proteins, or a metabolism made up of hundreds of proteins. For larger multicellular organisms like us, with slower generation times and smaller populations, the problem gets worse. Much worse.
So unless someone paved a highway to Mt. Whitney that went uphill every step of the way, Darwin’s engine would never get out of Death Valley. But a paved highway isn’t evolution, it’s design.
The hidden life of the cell, from the BBC. Stunning animation! Hat tip to JA for pointing to this.
Let the images speak for themselves.
Giving Evolution a New Meaning—Stylus
What do Han Chinese characters and proteins have in common? Both can be thought of as two- or three-dimensional functional shapes. Change their shapes significantly and you change what they mean or do.
For proteins the flow of functional information goes like this:
protein coding sequence ⇒ 3-D structure ⇒ function
If a code could be devised that specifies Han character traces in vector space, then the Han character set would have a similar information flow:
Han coding sequence ⇒ 2-D structure ⇒ meaning
And suddenly all the parallels between genetic code and language become a rich environment in which to test evolutionary scenarios.
This brilliant analogy between protein structure and Han characters occurred to Dr. Douglas Axe of the Biologic Institute several years ago. He and his collaborators have since devised a vector code analogous to the genetic code (see below) and implemented it in a program called Stylus. A paper describing its initial characteristics and making its source code available was published in the peer-reviewed journal PLoS One.
Figures from Stylus: A System for Evolutionary Experimentation Based on a Protein/Proteome Model with Non_Arbitrary Functional Constraints.
Dr. Axe and co-workers have now created a self-referential genome that describes in Han vector code the main elements of the program. Each sentence of the description is a unit of linked genes that specify a particular function, much like the tryptophan operon is a collection of genes that specify how to make tryptophan, all organized into a coordinated group so that their function is coordinated. This work has also been published in a peer-reviewed journal here.
Right now, the program is available as source code for people to experiment with. You can adjust various parameters that determine how much and what kind of genetic change takes place, how many generations you want the evolutionary experiment to run, etc. A web-based platform is the next goal. Let’s find out how much meaning random, undirected processes can create, shall we?
Why Stylus?
Stylus is an unique in silico model of gene→protein→function mapping in biology that provides a way to test evolutionary scenarios. This is because Stylus reflects the way mutation and selection work in biology.
Mutation (genetic causation) acts at a low level to produce changes to individual genes or segments of the genome. Selection (evolutionary causation) then evaluates the functionality of the resulting organism, i.e. the sum of all interactions between genes and proteins that affect the organism’s functional capability.
In Stylus, changes to the genome produce variations in the vector traces drawn by the program. Those traces are then compared to a standard Han character set to determine if their meaning is damaged or changed. Because the genome is linear and continuous, and includes sequences that control start and stop, domains that are used in multiple contexts, signals that control operon regulation etc., many of the kinds of sequence evolution proposed by biologists as a way to get new functional proteins or pathways can be tested in Stylus.
As the paper describing the Stylus genome states,
…the contribution of Stylus is to make evolutionary experimentation possible in a model world where low level genetic causation has the essential role that it has in the real world. Combined with the free Stylus software, the complete Stylus genome made freely available with this paper paves the way for analogy-based studies on a wide variety of important subjects, many of which are difficult to study by direct experimentation. Among these are the evolution of new protein folds by combining existing parts, the optimality and evolutionary optimization of the genetic code, the significance of selective thresholds for the origin and optimization of protein functions, and the reliability of methods used for homology detection and phylogenetic-tree construction.
(Source: biologicinstitute.org)
Stylus encodes a compact self-descriptive proteome
Stylus is an in silico model of the relationship between genetic sequence (genotype) and the ensemble of functional proteins encoded by the sequence (phenotype). The Stylus genotype specifies vector traces that reproduce Chinese Han characters. Those characters are then assembled into a functional proteome whose function is to describe key terms in the Stylus program (much like our genome specifies the protein building blocks necessary to asemble us.)
Shown above is a representation of the nine key terms of the Stylus proteome with their meanings in English. The bracketed terms present in abbreviated form how the genome and its proteome specify the genetic code. Although smaller than even the smallest known bacterial genomes at 70,000 bases, Stylus is similar in complexity in terms of numbers of proteins and protein domains encoded. Characters that are used multiple times are similarly colored, along with their meanings.
(Source: biologicinstitute.org)
Comparing Stylus with other computational models of in silico evolution
The key sequence to structure to function relationship found in biology is missing from nearly every other computational model of evolution, but present in Stylus. In addition, only Stylus uses a codon-based sequence to structure map like the one found in life.
For the peer-reviewed article describing the program see here.
(Source: biologicinstitute.org)
How do Chinese Han characters relate to proteins? Go to http://www.biologicinstitute.org/post/19298009298/introducing-stylus-new-software-for-a-new-take-on for an introduction to the idea. Then check out the peer-reviewed paper published in PLoS One.
← Older entries Page 1 of 11
