A new algorithm that simulates the evolution of complex biological systems may have some applications in the lab, according to a paper published in Nature Scientific Reports.
The Wolfram matha algorithm is based on a theory developed by scientists at the Massachusetts Institute of Technology and is the first time a computer-generated image has been used to represent complex systems.
The algorithm works by simulating the evolution over a number of generations of a given species, based on its genetic code.
Wolfram mathematician Paul Matzke, an assistant professor of mathematics at MIT, and graduate student Mark Borkowski used Wolfram matrices to represent the evolution, starting with the species A. melanogaster, which consists of two genomes and three copies of the protein coding for its own chloroplast (the first copy is the same as the chloropheretic cell’s own, and the second copy is from another species).
This first copy contains only the protein code for the chlorophyll, while the second and third copies contain genes coding for more complex proteins, like the chloroplasts, that make up the chlorobenzene that makes up chlorophyles.
This second copy also contains a gene for the pterosaurian egg-laying bird called Caudovorax, which has only one copy of the pteryosaurine gene, so this bird is represented as two cells with one egg.
The second and final copy of this bird, which is called the chloropteran, contains a different gene for its two feathers, which makes it a different species.
“By using this phylogenetic data, we can actually do the same thing with the animals we’re interested in,” Matzkelowski said.
“We can reconstruct phylogenies of organisms in the same way we reconstruct phylogenetic trees.”
The Wolframs algorithm is called a genome-wide genealogical model because it uses a tree that is represented by a list of genes and is used to predict the evolution.
The result is a list that represents a species’ evolutionary history in a much more efficient way than using a single phylogeny or even a simple tree.
Matzinksowski and Borkowski were able to predict a large number of the species’ phylogenies and even a whole family tree using this algorithm.
“It’s a great example of a method that’s useful for the genome-scale reconstruction of species evolution,” said lead author Michael D. Bostrom, a postdoctoral fellow in MIT’s Department of Mathematics.
“The fact that we were able a very, very large number and that we could predict the entire phylogeny from this means that this is really, really useful in the genome and across the whole animal kingdom.”
The results are based on the evolutionary history of more than 20,000 species and have been published in the journal Nature Scientific Studies.
Matzeski and Borrowowski hope that the Wolframs method can be extended to other types of data, such as those derived from the fossil record, which can help to predict species’ pasts.
“Our goal is to create a system that could represent the evolutionary histories of a whole range of complex organisms, not just just the vertebrates,” Matzesky said.
For example, Borrowowski and Matzeskowski believe that the tree of life, the set of all species, could be used to reconstruct a species tree that was able to estimate the species diversity of a particular environment.
“There are many more trees that could be built,” Borkowsky said.
The computational complexity of the Wolfumbers algorithm is not a problem for the fossil fossil record itself, Borkowksy said.
It has been applied in the past to tree models and other information systems.
“You can see a lot of data that are quite hard to reconstruct,” he said.
But it can also be used for other kinds of data.
“This is not only an extremely powerful way to reconstruct the fossil tree, it can be used as a tool to reconstruct animal phylogenies,” Borrowos said.
In the future, the team hopes to use Wolfram to model species that are not yet well-studied, such the species that lived in the Permian period, when the oceans were much more acidic.
“One of the things we are looking forward to is the possibility of reconstructing species that never existed,” Bajorowski said, “because there are some fossils of animals that were actually quite complex, or even extinct, at the time of the Perminian period.”
The research was funded by the National Science Foundation and the John D. and Catherine T. MacArthur Foundation.
The Harvard-MIT Joint Center for Artificial Intelligence is led by computer scientist John Huang, who is also an assistant director at the National Institute of Standards and Technology.
Additional co-authors of the paper are MIT graduate student and postdoctoral researcher Anand Swaminathan, and Harvard professor of computational biology Mark B. Löscher.
For more on the research, visit the Wolfams website at