Embracing Environmental Genomics and Machine Learning for Routine Biomonitoring

Abstract

Genomics is fast becoming a routine tool in medical diagnostics and cutting-edge biotechnologies. Yet, its use for environmental biomonitoring is still considered a futuristic ideal. Until now, environmental genomics was mainly used as a replacement of the burdensome morphological identification, to screen known morphologically distinguishable bioindicator taxa. While prokaryotic and eukaryotic microbial diversity is of key importance in ecosystem functioning, its implementation in biomonitoring programs is still largely unappreciated, mainly because of difficulties in identifying microbes and limited knowledge of their ecological functions. Here, we argue that the combination of massive environmental genomics microbial data with machine learning algorithms can be extremely powerful for biomonitoring programs and pave the way to fill important gaps in our understanding of microbial ecology.