Analysis of the Influence of Seasonal Water Column Dynamics on the Relationship between Marine Viruses and Microbial Food Web Components Using an Artificial Neural Network

Abstract

Artificial neural network analysis (ANN) is used to study the seasonal distribution of viruses and microbial food web (MFW) components in the open Adriatic Sea. The effect of viruses within the MFW is often overlooked, although viruses play an important role in microbial community dynamics. The results showed that the strongest influence is found in the nonlinear relationship between viruses and temperature. In addition, the algorithm showed that the number of viral populations in the P-limited open sea varies by season and according to the abundance of their main hosts, HB. A strong positive relationship between viruses and HB was found in more than 50% of the observed data. Moreover, this algorithm confirmed the association of the virus with the autotrophic part of the picoplankton and with heterotrophic nanoflagellates. The dynamics of the four resulting clusters, characterized by biological and environmental parameters, is described as a cyclic pattern in the water layer above the thermocline. Neural gas network analysis has been shown to be an excellent tool for describing changes in MFW components in the open Adriatic.