Caenorhabditis elegans has two sexes, hermaphrodites (XX) and males (XO). Males are very rare among the self-progeny of wild-type hermaphrodites (1:500 to 1:2000). Several genetic pathways are known to affect sex ratios, including meiotic chromosome segregation, dosage compensation and sex determination pathways. When inactivated by RNAi, over 100 genes were found to give the high incidence of males (him) phenotype (WormBase WS212). Yet, the functional mechanisms and pathway information for the majority of these genes remain unknown. Therefore, we decided to conduct a quantitative epistasis study to uncover the genetic pathways and networks of these him genes. To enable a high-throughput quantitative screen, we first developed a computer vision system to automatically measure C. elegans sex ratios. Using a motorized stage and a camera attached to the microscope, the system scans Petri plates and records images of animals. The system analyzes morphological features of the animals and outputs the numbers of adult males and hermaphrodites on each plate. To identify pairwise interactions among genes, we inactivate each gene by RNAi or mutation, and measure the sex ratio of these single mutants using the automatic phenotyping system. We use these single mutant data to compute an expected sex ratio for the double mutant assuming that the two genes function independently. We then inactivate two genes by applying RNAi on mutants and examine whether the observed double mutant sex ratio is significantly different than the expected value. An enhancing interaction is identified if the observed phenotype is more severe than the expected value, and a suppressing interaction is denoted if the observed phenotype is milder. In a pilot study, we selected mutants of five meiosis genes (brc-1, brd-1, him-3, him-5, and zhp-3) and screened them for possible interactions with 124 genes that gave the him phenotype. From over 600 gene pairs screened, we discovered over 100 genetic interactions. Among them, 55% interactions were suppressing, and 45% were enhancing. The majority (over 97%) of these interactions are novel ones that have not been reported before. These data suggest that our quantitative epistasis approach is a powerful tool to identify genetic interactions. 44 of the 124 him genes have homozygous viable mutants. We are now in the process of composing the 44x124 gene interaction map. We will report our results at the meeting.

Affiliation:
- Biochemistry and Cell Biology, Rice University, Houston, TX