Newbler v2

Sequences were assembled using Newbler v2.9 (Roche) with default parameters. For Illumina sequences, we used MEGAHIT v1.0.3 (Li et al., 2015 (link)) with meta-sensitive preset mode and minimum contig length of 180. Open reading frames (ORFs) were predicted from both contigs and reads using Prodigal v2.6.3 (Hyatt et al., 2010 (link)) in its metagenomics mode. Predicted proteins were then filtered to keep only non-redundant sequences of at least 60 amino acids length. ORFs from publicly available sequences were then clustered using CD-HIT v4.6 (Li et al., 2001 (link)) with 60% identity, 80% coverage and the following parameters: “-g 1 -n 4 -d 0” to give rise to the HVPC database.

The raw reads were quality checked and filtered out to remove low-quality reads as well as reads containing adapter sequences. Reads shorter than 60 bp were removed before performing the assembly. The resultant high-quality filtered reads were de novo assembled using the GS de novo assembler Newbler v2.9 (Roche, IN, USA) with the assembly parameters of a minimum overlap length of 40 bp and minimum overlap identity of 90%. The remaining unassembled reads (singletons) were extracted from the Newbler-trimmed 454 reads file using Seqtk software [73 (link)]. The CLC Genomics Workbench (http://www.clcbio.com (accessed on 12 August 2022)) was used to map back singletons against Newbler assembled transcripts, and unmapped singletons were eliminated as duplicates. The remaining reads were assembled using MIRA v4.2 with the following parameters: est, accurate, de-novo, 454. To remove duplicates, the singletons that failed to assemble by MIRA v4.2 were extracted and mapped back to the MIRA assembled transcripts using CLC Genomics Workbench. Further, transcripts less than 200 bps generated by MIRA v4.2 were removed, and the remaining transcripts were merged with the transcripts generated by Newbler v2.9 to get the final transcriptome assembly. The graphical representation of the detailed methodology followed is shown in Figure 8.

Illumina reads that passed quality control were assembled, each sample separately, using megahit v1.1.1 (Liu et al., 2015 (link)) with the parameters (-t 20 -m 0.8 --preset meta-sensitive), whereas 454 reads were assembled using Newbler v.2.9 (Roche) using default parameters, since Newbler has better performance than other assemblers for Roche 454 reads (Kumar and Blaxter, 2010 (link)). Produced contigs were renamed according to the following pattern: freshwater_SRRAccession.contig000XXXXXX, where SRRAccession is the SRA Run accession number of the particular sample used for assembly and XXXXXX is a serial number. Metagenomes that produced no contigs longer than or equal to 1,000 bp were removed from further analysis.