Dictyostelium Aggregation

This example describes a pathway that regulates the transition from growth to development in Dictyostelium development. It illustrates the integration of prior knowledge and of data from non-genetic experiments into network construction. It also illustrates the analysis of parallel and converging pathways. pkaC encodes the catalytic subunit of the cAMP-dependent protein kinase PkaC. pkaR encodes the regulatory subunit which inhibits PkaC in the absence of cAMP (that information was entered as prior knowledge). acaA encodes the adenylyl cyclase enzyme which produces cAMP from ATP and regA encodes a phosphodiesterase enzyme which degrades cAMP. The relationships between regA and the other genes were published in Shaulsky, G., Fuller, D., and Loomis, W. F. (1998). A cAMP-phosphodiesterase controls PKA-dependent differentiation. Development 125, 691-699. pufA encodes an RNA-binding protein which inhibits the translation of PkaC. That information was derived from biochemical studies and was therefore entered as prior knowledge. yakA encodes a protein kinase that inhibits the function of pufA indirectly. The relationships between yakA, pufA and the pkaC network were published in Souza, G. M., da Silva, A. M., and Kuspa, A. (1999). Starvation promotes Dictyostelium development by relieving PufA inhibition of PKA translation through the YakA kinase pathway. Development 126, 3263-3274 and in Souza, G. M., Lu, S., and Kuspa, A. (1998). YakA, a protein kinase required for the transition from growth to development in Dictyostelium. Development 125, 2291-2302. This example has been discussed in detail in Zupan, B., Demsar, J., Bratko, I., Juvan, P., Halter, J. A., Kuspa, A., and Shaulsky, G. (2003). GenePath: a system for automated construction of genetic networks from mutant data. Bioinformatics 19, 383-389.

Dictyostelium Sporulation

This is an example of a pathway that regulates terminal differentiation (spore formation) in Dictyostelium. It illustrates the flexibility of phenotypic definitions in GenePath, the discovery of parallel pathways and the integration of findings from genetic experiments and prior knowledge during network construction. pkaC encodes the catalytic subunit of the cAMP-dependent protein kinase PkaC. pkaR encodes the regulatory subunit which inhibits PkaC in the absence of cAMP (that information was entered as prior knowledge). regA encodes a phosphodiesterase enzyme which degrades cAMP. The relationships between regA and the other genes were published in Shaulsky, G., Fuller, D., and Loomis, W. F. (1998). A cAMP-phosphodiesterase controls PKA-dependent differentiation. Development 125, 691-699. tagB and tagC encode composite proteins that carry serine protease domains and ATP-dependent membrane transporters of the MDR family. The findings on tagB and tagC have been published in Shaulsky, G., Escalante, R., and Loomis, W. F. (1996). Developmental signal transduction pathways uncovered by genetic suppressors. Proc Natl Acad Sci USA 93, 15260-15265 and in Shaulsky, G., Kuspa, A., and Loomis, W. F. (1995). A multidrug resistance transporter serine protease gene is required for prestalk specialization in Dictyostelium. Genes Devel 9, 1111-1122. dhkA encodes a hybrid histidine kinase protein of the two-component system family. The data on dhkA are from Wang, N., Shaulsky, G., Escalante, R., and Loomis, W. F. (1996). A two-component histidine kinase gene that functions in Dictyostelium development. EMBO J 15, 3890-3898 and Wang, N., Soderbom, F., Anjard, C., Shaulsky, G., and Loomis, W. F. (1999). SDF-2 induction of terminal differentiation in Dictyostelium discoideum is mediated by the membrane-spanning sensor kinase DhkA. Mol Cell Biol 19, 4750-4756.

Programmed Cell Death

Programmed cell death of the nematode C. elegans from the study by Metzstein et al. (TIG, Oct 1998) that includes four genes (egl1, ced3, ced4, ced9) and five data sets (cell viability, cell death, cell killing, ced3-induced killing, and egl1-induced killing). Cell death, cell killing and ced3-induced killing were merged into a single phenotype (programmed cell death), for which GenePath discovers the genetic network that is consistent with the one presented by Metzstein et al.

Program Cell Death (viability)

Programmed cell death of the nematode C. elegans from the study by Metzstein et al. (TIG, Oct 1998) that includes four genes (egl1, ced3, ced4, ced9) and five data sets (cell viability, cell death, cell killing, ced3-induced killing, and egl1-induced killing). Cell death, cell killing and ced3-induced killing were merged into a single phenotype (programmed cell death), for which GenePath discovers the genetic network that is consistent with the one presented by Metzstein et al.

Double-Blind Experiment

In most cases, we were aware of the genetic network that GenePath was supposed to infer. To test GenePath more rigorously, we used a double-blind schema where one of the authors (A.K.) selected a published genetic problem (Dauer larva formation in C. elegans, Riddle, D. L., Swanson, M. M., and Albert, P. S. (1981). Interacting genes in nematode dauer larva formation. Nature 290, 668-671), coded the gene and phenotype names and gave the data with no indication on the nature of experiments to the other authors who used GenePath to obtain the resulting network. GenePath performed its task and produced a network that was identical to the published one with one exception (gene 2) where the authors of the original paper did not follow the rules of epistasis consistently. We entered all the experiments that include gene 2 but instructed the program to ignore them. You may reset the program to include the experiments in the ‘experimental data’ window. It should be noted that the pathway presented here and in the original paper is not considered to be correct today, probably because some of the mutations were partial (the rules we use are best suited for null mutations) and also because the authors used temperature shifts that may have confounded their findings. This is an important illustration of how GenePath is limited by the quality of the genetic data. This example has been discussed in detail in Zupan, B., Demsar, J., Bratko, I., Juvan, P., Halter, J. A., Kuspa, A., and Shaulsky, G. (2003). GenePath: a system for automated construction of genetic networks from mutant data. Bioinformatics 19, 383-389.

Dictyostelium Communication and Adhesion

In this example, GenePath considers two types of experiments. One type involves the analysis of single and double mutations and the other examines the effect of one gene on the expression of another. This example also illustrates the use of the ‘Notebook’ option. lagC, lagD and comC are genes that were cloned in a screen for communication mutants in Dictyostelium discoideum. Knocking out any one of these genes leads to defects in development but developing them in chimera with wild type cells can rescue the defects, indicating that the genes are involved in signal production. Chimerae of any two mutants also fail to develop, indicating that the three genes are involved in one signaling pathway. Since lagC has been described before as an adhesion gene, we think that the pathway regulates cell-cell communication via adhesion. The data given here are from: Kibler, K., Svetz, J., Nguyen, T. L., Shaw, C., and Shaulsky, G. (2003). A Cell-Adhesion Pathway Regulates Intercellular Communication During Dictyostelium. Dev Biol 264, 506-521.