Proceeding talk – Theme: Genes.
Abstract
Binning metagenomics reads remains a crucial step in metagenomics analysis. In this work we presented MetaProb, an assembly-assisted approach for reference-free metagenomic binning. Our approach can deal with short and long reads in a novel probabilistic framework, by using probabilistic sequence signatures. We compared the binning performances over several short and long reads datasets against other state-of-art binning algorithms, showing that MetaProb achieves in most cases the best performances in terms of F-measure. The estimation of the number of species in a metagenomic sample can be performed with MetaProb, adding a degree of freedom in the analysis. On a real fecal metagenomic data MetaProb was able to detect the most abundant species with high precision.
Authors
Samuele Girotto, Dept. Information Engineering, University of Padova, Italy
Cinzia Pizzi, Dipartimento di Ingegneria dell’Informazione, Univ. of Padova, Italy
Matteo Comin, Dept. Information Engineering, University of Padova, Italy