Genetic Algorithms use Darwinian principles to solve mathematical (programming) problems. Are they actually useful in obtaining knowledge in the opposite direction, i.e, in Evolutionary Biology.
Yes, genetic algorithms have contributed to evolutionary biology.
First, by being applied to optimisation problems in evolutionary biology, e.g. reconstructing phylogenetic trees [http://molevol.lysine.umiacs.umd.edu/molevolfiles/garli/zwicklDissertation.pdf]
Second, by simulating biological systems, e.g. analysing how biological sequences evolve over time, with applications in drug discovery [http://www.ncbi.nlm.nih.gov/pubmed/23179493]
Third, by contributing general knowledge about evolvability, e.g. evolutionary consequences of redundancy, genome organisation, epistasis etc.
In the last two years, I have been teaching and learning Genetic Algorithms (GAs) for applications in solid state physics, i.e., by obtaining new structures of the so called new materials. The ideas and concepts behind this algorithm sounds interesting and are giving excellent results in this field. However, concerning your question, my experience tells me that first, Evolutionary Biology can (and should) improve the GA by, as an example, fixing the rules of mutation rates, parent selections and/or crossover tasks. Better saying, it can improve what we call "boundary conditions" in Physics (and I think it is necessary). After that, it will be naturally easy to predict that GAs can be useful to obtain knowledge in the opposite direction, as described in your question. From my point of view, there is still a lot of work to do in GAs in order to be capable to make "reverse engineering" in the Evolutionary Biology.
Yes, genetic algorithms have contributed to evolutionary biology.
First, by being applied to optimisation problems in evolutionary biology, e.g. reconstructing phylogenetic trees [http://molevol.lysine.umiacs.umd.edu/molevolfiles/garli/zwicklDissertation.pdf]
Second, by simulating biological systems, e.g. analysing how biological sequences evolve over time, with applications in drug discovery [http://www.ncbi.nlm.nih.gov/pubmed/23179493]
Third, by contributing general knowledge about evolvability, e.g. evolutionary consequences of redundancy, genome organisation, epistasis etc.
Yes, in a way. You can't really simulate an real world evolutionary system using an Evolutionary Algorithm. But an idea of how something can evolve can be tested using a simulation.
I personally coauthored something like that:
http://dl.acm.org/citation.cfm?id=2482750
GARM is a cool application of genetic algorithms in landscape genetics:
http://cel.dbs.umt.edu/cms/index.php/software/garm
.
A couple more examples in response to your question ( and the short answer is yes):
http://www.bioinfo.de/isb/2004040024/main.html
http://warnercnr.colostate.edu/~gwhite/fw663/lebreton.pdf
My research team applied genetic algorithms for multiple sequence alignments of DNA
- Badges
- Science topic
- Similar topics
- Biological Science
- Evolutionary Biology