Recursive expectation-maximization clustering: A method for identifying buffering mechanisms composed of phenomic modules

Abstract

Interactions between genetic and/or environmental factors are ubiquitous, affecting the phenotypes of organisms in complex ways. Knowledge about such interactions is becoming rate-limiting for our understanding of human disease and other biological phenomena. Phenomics refers to the integrative analysis of how all genes contribute to phenotype variation, entailing genome and organism level information. A systems biology view of gene interactions is critical for phenomics. Unfortunately the problem is intractable in humans; however, it can be addressed in simpler genetic model systems. Our research group has focused on the concept of genetic buffering of phenotypic variation, in studies employing the single-cell eukaryotic organism, S. cerevisiae. We have developed a methodology, quantitative high throughput cellular phenotyping ͑Q-HTCP͒, for highresolution measurements of gene-gene and gene-environment interactions on a genome-wide scale. Q-HTCP is being applied to the complete set of S. cerevisiae gene deletion strains, a unique resource for systematically mapping gene interactions. Genetic buffering is the idea that comprehensive and quantitative knowledge about how genes interact with respect to phenotypes will lead to an appreciation of how genes and pathways are functionally connected at a systems level to maintain homeostasis. However, extracting biologically useful information from Q-HTCP data is challenging, due to the multidimensional and nonlinear nature of gene interactions, together with a relative lack of prior biological information. Here we describe a new approach for mining quantitative genetic interaction data called recursive expectation-maximization clustering ͑REMc͒. We developed REMc to help discover phenomic modules, defined as sets of genes with similar patterns of interaction across a series of genetic or environmental perturbations. Such modules are reflective of buffering mechanisms, i.e., genes that play a related role in the maintenance of physiological homeostasis. To develop the method, 297 gene deletion strains were selected based on gene-drug interactions with hydroxyurea, an inhibitor of ribonucleotide reductase enzyme activity, which is critical for DNA synthesis. To partition the gene functions, these 297 deletion strains were challenged with growth inhibitory drugs known to target different genes and cellular pathways. Q-HTCP-derived growth curves were used to quantify all gene interactions, and the data were used to test the performance of REMc. Fundamental advantages of REMc include objective assessment of total number of clusters and assignment to each cluster a log-likelihood value, which can be considered an indicator of statistical quality of clusters. To assess the biological quality of clusters, we developed a method called gene ontology information divergence z-score ͑GOid_z͒. GOid_z summarizes total enrichment of GO attributes within individual clusters. Using these and other criteria, we compared the performance of REMc to hierarchical and K-means clustering. The ma...

A humanized yeast phenomic model of deoxycytidine kinase to predict genetic buffering of nucleoside analog cytotoxicity

“…Recursive expectation-maximization clustering ( REMc ) was used to identify modules of genes that shared similar profiles of buffering or promoting nucleoside toxicity of gemcitabine or cytarabine [40] (see Fig. 3A-F ; Table 1 ; Additional File 5 ).…”

“…As described previously, REMc results were assessed with GO Term Finder for Gene Ontology functional enrichment [41] and heatmaps generated by first adding data regarding the main effect of the gene knockout or knockdown ( i.e. , no drug) on cell proliferation, termed ‘shift’ (see methods), followed by hierarchical clustering [40,41]. GO Term Average ( GTA ) scores, which are based on the average and standard deviation of drug-gene interaction for all genes of each GO term [39], were used as a complement to REMc/GTF for identifying functions that buffer or promote drug effects (Table 2, Fig.…”

Phenomic Assessment of Genetic Buffering by Kinetic Analysis of Cell Arrays

“…To investigate genetic buffering networks, mutant strain cell arrays can be challenged with drugs or environmental variables and/or systematically combined with mutations of interest (3–8). Rigorous quantification of cell proliferation phenotypes enables accurate classification of gene interactions based on their strength of effect, thereby facilitating identification of genetic modules and buffering networks (3, 9, 10). Development of quantitative cell array phenotyping for growth curve analysis is in a relatively early stage (11), as its importance came to light only after work with the yeast mutant arrays, constructed fairly recently (12), revealed a high frequency of gene interaction (3, 6, 13, 14).…”

“…Gene interaction varies as a function of cellular context, and thus the interpretation of gene interaction networks depends on the biological nature of the aggregated phenotypic data (9). Q-HTCP can be used to obtain gene interaction profiles for any drug, environmental factor, or gene of interest.…”

Introduction to Focus Issue: Genetic Interactions

“…The contribution by Carter et al 40 highlights the relevance of information theory to tease out the differences between genetic modules and classically defined pathways. The work by Guo et al 41 introduces a new recursive maximum-likelihood approach to partition an interaction network into modules, and applies it to a gene-drug interaction data set. Both papers emphasize that research beyond classical clustering algorithms is necessary in order to understand how genetic interactions are organized, and how to gain novel biological and medical insight out of them.…”