Universal adapter sequence

Ribosomal RNA reads were first removed from metatranscriptomes using SortMeRNA v2.1 [115 (link)] with bacterial, archaeal and eukaryotic rRNA databases, and results were passed to repair.sh (BBMap [116 ]) to regenerate paired-end reads. Metagenomic and non-ribosomal metatranscriptomic read quality was inspected using FastQC [117 ]. Reads were trimmed using Trimmomatic v0.38 [118 (link)] with a minimum Phred score of 30 and adapter sequence removal via the ILLUMINACLIP option with Illumina universal adapter sequence (AGATCGGAAGAG). Reads were also trimmed to a maximum of 240 bp (DNA) and 115 bp (RNA) long to remove additional low-quality bases, and reads <80 bp (DNA) and <60 bp (RNA) long were discarded. A sediment sample from site 9, replicate 1 (S9R1), was subjected to additional trimming (headcrop 10) due to low base quality.

Quality of the sequenced files was checked using FastQC tool (0.11.9) (Andrews, 2010 ), followed by removal of low quality bases (–nextseq-trim, Q < 20), Illumina Universal adapter sequence and reads with less than 30-bp length using Cutadapt (2.10) (Martin, 2011 (link)). To access the quantity of host genomic DNA and other contaminants, Kraken (2.0.9-beta) (Wood et al., 2019 ) was used, and the reports were summarized using Krona (2.7.1) (Ondov et al., 2011 ). All the files were then aligned to human genome (assembly version GRCh38) using HISAT2 (2.2.0) (Kim et al., 2015 (link)), and unmapped reads were extracted using SAMTOOLS (1.10) (Li et al., 2009 (link)) and converted to FASTQ format using BEDTOOLS (2.29.2) (Quinlan and Hall, 2010 (link)) bamToFastq option.