Project description
Improved binning improves the quality of metagenome-assembled genomes
Human and environmental microbiomes are diverse microbial communities that play essential roles in human health and ecosystem functioning. Our understanding of their genes, products and dynamic interactions has increased with the use of metagenomics, the study of the structure and function of entire nucleotide sequences obtained from all the organisms in a sample. However, the quality of the metagenome-assembled genomes (MAGs) reconstructed from metagenome data suffers from challenges with binning, which is the grouping of sequences into species-wise bins. With the support of the Marie Skłodowska-Curie Actions programme, the Metagenome binning project plans to address the problems with a novel binning algorithm that improves the binning of low-abundance and highly conserved sequences.
Objective
Metagenome-assembled genomes (MAGs) obtained from metagenomics are of fundamental value to understanding diverse ecological niches of microbes such as the human gut, with applications in medicine, biotechnology, and climate science. However, the quality of MAGs constructed with state-of-the-art tools is often unsatisfactory and worse than the self-reported quality. The main source of error is binning, a computational step that groups sequences assembled from short sequencing reads (contigs) into species-wise bins. The two chief challenges are accurately binning (1) genomes with low abundance and (2) highly conserved regions. Due to cross-mapping of reads, the contigs from conserved regions appear to have abundances equal to the sum of the abundances of the related species or strains. As conventional binning tools all rely on clustering contigs according to their abundances across samples, conserved regions end up forming separate bins. Besides, most existing methods optimise quality measures (purity and completeness based on conserved marker genes) and assess the final quality on these very measures, leading to highly optimistic results. I aim to solve these problems by developing a binning algorithm that applies
i) linear mixture models using non-negative matrix factorization to account for cross-mapping,
ii) Poisson statistics to accurately model low abundance, and
iii) Bayesian statistics-based multinomial clustering to calculate bin numbers. Importantly, it does not require marker gene-based quality measures for binning.
By improving the binning of low-abundance and highly conserved contigs, this approach should yield more high-quality MAGs, thereby enhancing a multitude of downstream metagenomic analyses for all areas of microbiome research.
Keywords
Programme(s)
- HORIZON.1.2 - Marie Skłodowska-Curie Actions (MSCA) Main Programme
Funding Scheme
HORIZON-TMA-MSCA-PF-EF - HORIZON TMA MSCA Postdoctoral Fellowships - European FellowshipsCoordinator
80539 Munchen
Germany