Agnostic Tools for Life Detection: A Multiscale Approach to Biological and Chemical Reaction Networks

Life manifests itself across a variety of different operative scales, ranging from molecular networks up to ecologies. Deep conceptual difficulties arise when trying to define exactly what lifes distinguishing attributes are, and how they may be quantitatively distinguished from abiotic systems that are complicated but demonstrably not alive. Here we assembled different biotic and abiotic chemical reaction networks to test whether network-level (macroscale) and reaction-level (microscale) metrics can impartially distinguish biotic from abiotic category networks. Macroscale attributes such as statistical tests for a discrete networks heavy-tailed connectivity distribution generally distinguish bona fide biotic and protobiotic networks (E. coli heterotrophy, pruned protometabolism, KEGG) from abiotic networks but with two significant exceptions (radiolytic network and the Open Reaction Database). Microscale attributes such as an analysis of the frequency of a set of six primitive reaction motifs shows a depletion of some motifs in all chemical networks and the prevalence of others in specific networks; all networks were readily distinguishable from random network variants across all motifs. A comparison of the chemical spaces spanned by the different networks points to similarity and dissimilarity relationships between networks. Independent component analysis of an aggregate of all measured attributes, across all contexts, reliably distinguishes abiotic, prebiotic and biotic networks from one another. The combination of macro- and microscale attributes forms an agnostic toolset that may help to evaluate the prebiotic plausibility of different candidate settings for the origins of life, and may inform new ways of detecting and recognizing living systems (engineered or extraterrestrial) that differ from Terran biochemistry.