Systematic analysis of pleiotropy in C. elegans early embryogenesis

Abstract

Pleiotropy refers to the phenomenon in which a single gene controls several distinct, and seemingly unrelated, phenotypic effects. We use C. elegans early embryogenesis as a model to conduct systematic studies of pleiotropy. We analyze high-throughput RNA interference (RNAi) data from C. elegans and identify "phenotypic signatures", which are sets of cellular defects indicative of certain biological functions. By matching phenotypic profiles to our identified signatures, we assign genes with complex phenotypic profiles to multiple functional classes. Overall, we observe that pleiotropy occurs extensively among genes involved in early embryogenesis, and a small proportion of these genes are highly pleiotropic. We hypothesize that genes involved in early embryogenesis are organized into partially overlapping functional modules, and that pleiotropic genes represent "connectors" between these modules. In support of this hypothesis, we find that highly pleiotropic genes tend to reside in central positions in protein-protein interaction networks, suggesting that pleiotropic genes act as connecting points between different protein complexes or pathways.

Figure 1. The distribution of the percentage of genes against the number of defects in their phenotypic profiles.

We plot the distribution of the percentage of genes against the number of defects (brown bars) and compare with that of randomly permuted datasets (blue bars). The error bars show the standard deviation of the percentages of genes in the randomly permuted datasets. On average, genes in the dataset show 7 cellular defects in their phenotypic profiles. About 10% of the genes show 15 or more defects, much higher than that of the randomly permuted dataset.

Figure 2. A correlation map for pairs of defects involved in C. elegans early embryogenesis.

We calculate the ratio of observed co-occurrence to the expected co-occurrence for every pair-wise combination of defects and plot the ratios into a correlation map. A ratio that is higher than 1 indicates the two defects are more likely to co-occur than expected by chance. Some defects, such as P₁/AB nuclear separation—cross-eyed and four-cell stage nuclei—size/shape (pointed to with a black arrow), co-occur at a very high frequency. In this map, the co-occurring defects are grouped together by hierarchical clustering.

Figure 3. LPCA analysis of the cellular defects.

The X axis and Y axis represent the first and the second principal components in LPCA analysis, respectively. The data points labeled with numbers represent the cellular defects, which can be separated according to first two components. The correspondence between numbers and defects are the same as in Figure 2. The defects in proximity in the graph are likely to be closely related biologically.

Figure 4. A scheme of our method for identifying phenotypic signatures.

We start with pre-defined classes and search for defects that are enriched in each class. The enriched defects compose the phenotypic signature of a given class. We then search for genes that were originally not included in the class but can be matched with the phenotypic signature the class. In this process, a gene may be assigned to multiple classes. The defects shown in orange and green represent phenotypic signatures of two different classes. In this example, Gene D is re-assigned to both classes.

Figure 5. The phenotypic signature for the cell polarity class.

This signature consists of 7 defects (highlighted in red). One additional gene, K09H11.3 (RGA-3) (pointed to by a red arrow), is assigned to the cell polarity class.

Figure 6. The phenotypic signature for the chromosome function class.

This signature consists of 6 defects (highlighted in red). Eight additional genes (pointed to by red arrows) are assigned to the chromosome function class.

Figure 7. Distribution of pleiotropy indices.

We define Pleiotropy Index (PI) as the number of functional classes a gene is assigned to. We plot the distribution of genes with different PIs in a pie chart. We observe that more than half of the genes are pleiotropic (PI≥2), and only 3% of the genes are highly pleiotropic (PI≥5).

Figure 8. Pleiotropic genes as “module connectors”.

The extensive existence of pleiotropic genes suggests that gene modules are overlapping rather than separate from one another. Genes assigned to the same functional class are represented as a module and pleiotropic genes correspond to the intersections of modules. In the illustrated example, the most pleiotropic kinases, including dom-6, mpk-1, plk-1 and air-1, connect most of the modules in early embryogenesis into a “module network”.

Figure 9. Scatter plot of Relative Pleiotropy Score and rank of betweenness.

The rank of betweenness is significantly correlated with the Relative Pleiotropy Score. This correlation is largely contributed by the highly pleiotropic genes (upper right corner).