Contig-based taxonomical and functional classifications.

The assembled contigs were annotated using an integrated pipeline MetaErg 1.2.0 (99). MetaErg performed HMMs profile similarity searches against several databases, including Pfam-A (100), TIGRFAM (101), FOAM (102), metabolic-hmms (103), and casgenes.hmm (104). MetaErg also performed DIAMOND (double index alignment of next-generation sequencing data) searches against Swiss-Prot (105) and the MetaErg in-built database GenomeDB. Mapping files generated from searches against Swiss-Prot, FOAM, and TIGRFAMs databases were incorporated in MinPath (106) to infer to KEGG (107) and MetaCyc (108) metabolic pathways. To weigh the relative abundances of the taxonomic, functional, and pathway compositions of the metagenome samples, individual BAM files were generated using BBMap (95) by aligning the metagenome reads of each sample to the coassembled contigs. A depth file was constructed with the jgi_summarize_bam_contig_depths script in MetaBat 2 (109) using the BAM files.