Research Associate
kjkim@u.washington.edu
I am
building computational models of genetic regulatory networks. These models reconstitute the biological interactions between the core regulatory genes (and their products) in a system of equations or rules. Though the interactions and
properties of genes are often simple, the behavior of the networks is not.
I am involved in several research projects described here:
Building models of genetic networks:
- Constructing a model of the Drosophila endocycle: The endocycle is an abbreviated cell cycle where cells replicate their DNA but do not divide, resulting in cells with amplified DNA content. The goal of the research is to understand how this network
can be modulated to produce faster or slower cycling, which translates to changes in the rate of growth. I am collaborating with the laboratory of Bruce Edgar at the Fred Hutchinson Cancer Research Center, and with George von Dassow at the CCD.
- Constructing a model of the C. elegans vulva induction network: During larval development in C. elegans, a line of 6 cells can each be induced to assume one of 3 cell fates. Normally, a nearby anchor cell releases a morphogen that induces the closest cell to assume a 1' fate. This cell in turn induces its two neighbors to assume a 2' fate through Notch/Delta signaling, and the other cells remain uninduced. This patterning is essential to formation of a functional vulva, and the networks responsible appear to be highly robust to variation. This project is a collaboration with the laboratory of Marie-Anne Felix at Institut Jacques-Monod, in Paris, and with Erika Hoyos, Ed Munro, and Eli Meir at the CCD.
Simulating evolution in models of genetic networks:
- Many genetic networks are robust to perturbations due to environmental stress and genetic mutation; and one could imagine that such robustness may be selected for by evolution. To explore this, I am simulating (along with Vilaiwan Fernandes) mutation and selection in the segment polarity and endocycle networks. We simulate a population of individuals possessing a network, allow mutations to alter the quantitative behavior of the network, and select for those individuals that produce the correct phenotype (network output). In this simulation, we can explore the effects of population size, ploidy, mutation rate, and reproduction mode (sexual vs. asexual) to determine how it influences evolution of robustness. We are also exploring simulating speciation by testing the simulated genetic compatibility between reproductively isolated populations.
Building software tools to investigate genetic networks:
- The computational studies above are driven by biological questions, but also require the creation of software tools to facilitate their study. I am the current maintainer of the gene network simulation software Ingeneue, created by Ed Munro, Eli Meir, George von Dassow and Garrett Odell, with additional programming by Bill Sunderland. I also have written a package in Mathematica that allows simulation of Ingeneue networks, and am developing software to automate and efficiently simulate evolution of genetic networks within finite populations of simulated individuals. This package will allow simulation of evolution in any Ingeneue network.