Detecting Seasonal Marine Microbial Communities with Symmetrical Non-negative Matrix Factorization

With the development of high-throughput and low-cost sequencing technology, a large amount of marine microbial sequences is generated. So, it is possible to research more uncultivated marine microbes. The marine microbial diversity, the association patterns among marine microbial species and environment factors are hidden in these large amount sequences. Understanding these association patterns has a high potential for exploiting the marine resources. Yet, very few marine microbial association patterns are well characterized even with the weight of research effort presently devoted to this field. In this paper, with the 16S rRNA tag pyrosequencing data taken monthly over 6 years at a temperate marine coastal sits in West English Channel, we first introduced a neighbor-seeds based heuristic clustering method called as NbHCluster by incorporating an adaptive neighbor set expanding procedure and a greedy heuristic clustering procedure, to generate the operational taxonomic units (OTUs), and utilized the mutual information (MI) algorithm to construct the spring, summer, fall, and winter seasonal marine association networks of microbe and environmental factors. Then, we used the fuzzy clustering framework by defining a clique-node similarity matrix and adopting the symmetrical non-negative matrix factorization method, to detect the association community patterns and structures in the four seasonal marine networks. The results show that the four seasonal marine microbial association networks have characters of complex networks, and the marine microbial association patterns are related with the seasonal variability; the same environmental factor influence different species in the four seasons; and the correlative relationships are stronger between OTUs (taxa) than with environmental factors.