Metagenome, Resistome, and Replicome for Causal Inferencing

(formerly RAPToR)


Principal Investigator: Prof. Giri Narasimhan
Principal Architect: Vitalii Stebliankin
Other Contributors: Camilo Valdes, Musfiqur Sazal, Kalai Mathee



In this work, we present a pipeline for metagenomic analysis called MeRRCI for distributed computation of (meta)resistome, microbial composition, and bacterial replication rates (metareplicome), followed by Bayesian network analysis to discern causal relationships within the microbiome. First, the Map-Reduce procedure is used to map the metagenomic sequence reads against the collection of reference genomes (RefSeq) and resistance genes (CARD). Second, the read coverage pattern is used to compute microbial composition profile, metareplicome, and metaresistome. The replication is measured with Peak to Trough Ratio (PTR) method, which reflects the mechanism of bacterial cell division (more reads are observed at the origin of replication when bacteria are in a phase of active growth). Finally, the computed variables are used to construct a causal Bayesian network that defines the joined probability distribution of the multivariate system. The resulting causal network aims to highlight the most significant associations within the microbiome and rule out the coincidental correlation by exploiting the conditional independence information between the variables of interest.
Download Github Site: MeRRCI
Contact: Prof. Giri Narasimhan



Vitalii Stebliankin, Musfiqur Sazal, Camilo Valdes, Kalai Mathee, and Giri Narasimhan. A novel approach for combining metagenome, resistome, replicome, and causal inference to determine microbial survival strategies against antibiotics (Under Review, 2022)


Vitalii Stebliankin