Supervised classification of metatranscriptomic reads reveals the existence of light-dark oscillations during infection of phytoplankton by viruses

In the era of next generation sequencing technologies microbial species identification is typically performed using sequence similarity and sequence phylogeny based approaches. Particularly challenging is the discrimination of closely related sequences such as auxiliary metabolic genes (AMGs) in cyanobacteria and their viruses (cyanophages). Here we developed a method which combines Support Vector Machine based classification of AMGs short fragments and Empirical Mode Decomposition of periodic features in time-series. We applied this method to investigate the transcriptional dynamics of viral infection in the ocean, using data extracted from a previously published metatranscriptome profile of a naturally occurring oceanic bacterial assemblage sampled Lagrangially over 3 days. We discovered the existence of light-dark oscillations in the expression patterns of AMGs in cyanophages which follow the harmonic diel transcription of both oxygenic photoautotrophic and heterotrophic members of the community. These findings suggest that viral infection might provide the link between light-dark oscillations of microbial populations in the North Pacific Subtropical Gyre.