Raw reads were filtered using the Kneaddata pipeline v0.7.7 (https://github.com/biobakery/kneaddata) where human contaminations and low-quality reads were filtered before constructing the initial assembly via metaSPAdes v3.13.0 with default parameters. The binning, bin refinement, bin reassembly, and initial bin taxonomy classification were then performed using the metaWRAP pipeline v1.3 with the quality threshold for the bins set at >50% completeness and <10% contamination rate. The taxonomy of the final bins was further assessed using GTDB-Tk v2.1.1 using the Genome Taxonomy Database release 214. MAGs were accessed for completeness and contamination using CheckM v1.2.2