A novel phylogeny-based pattern selection algorithm and its application to microbiomic data

Discriminative patterns describe significant differences between different types of subjects, and often provide insights to critical properties of the problem at hand. Pattern-based classifiers can directly utilize discriminative patterns to predict unseen samples by a majority voting or aggregation mechanism. Therefore, we are concerned with not only finding useful individual patterns, but also the effectiveness of the pattern set as a whole; and it is imperative to ensure the relevancy and non-redundancy of the discriminative patterns. Few studies have evaluated pattern redundancy via examining samples covered by the patterns; and in those that do, the focus has been mostly on the proportion of overlapping samples, suggesting that a great deal of information on non-overlapping samples was overlooked. To address this issue, we present a novel pattern selection algorithm that estimates pattern redundancy by not only the proportion of overlapping samples, but also the resemblance of non-overlapping samples. The proposed method was applied on two real microbiomic datasets, with the aim of providing new insights on the interactions between microbes, and their effects on the host. When compared with other robust classifiers and feature selection heuristics, our pattern selection algorithm led to diverse and compact sets of final patterns that demonstrated comparable or even superior predictive capabilities.