CoPR: Collective Pattern Recognition-a Framework for Microbial Community Activity Analysis

BackgroundMicrobial community activities provide essential information on understanding bacterial communities. Unfortunately, they are generally not directly observable. We rely on longitudinal abundance profiles to get insight into microbial community activities. Often datasets do not have sufficient longitudinal sampling points to successfully apply our algorithms. Hence, in this paper, we are interested in analysing multiple datasets from similar environments to alleviate the aforementioned problem. Furthermore, we wish to see whether collective pattern recognition would enhance our understanding of microbial community activities. ResultsIn this paper, we present CoPR, a framework for collective microbial longitudinal abundance data. Our visualisation shows that a single pattern for temporal abundance variation does not exist. However, it also indicates that even complete individuality does not exist. Consequently, our visualisation highlights the individuality and conformity in the temporal variation of abundance profiles of similar host environments. We also identify different characteristics in the TVAP (Temporal Variation of Abundance Profile) patterns with regards to cohesion and separation. ConclusionsCoPR helps gain essential insights into the microbial communities and their heterogeneity through visualisation tools. This paper also highlights the choice between individuality and conformity in microbial community data analysis.