We used custom PERL scripts to retrieve the index information and identify and trim the amplification primer sequences. We discarded any DNA sequence shorter than 50 bp, with the exception of sequences amplified with plant and bryophyte primers for which short amplification products were expected (see Supplemental Table 1 ). We also discarded any trimmed read pair for which the difference in sequence length was greater than 5 bp between the two paired-end reads to eliminate any reads where primers were not found in both reads. We then merged the paired reads into a single consensus DNA sequence using PANDAseq with default parameters26 (link). We used Mothur27 (link) to cluster unique DNA sequences and counted how many reads carried each unique DNA sequence.
While all raw sequences are freely available online (accession number SRP058316), we only describe here the analyses of macroorganism DNA sequences for sake of simplicity (microorganism sequences can be easily analyzed using standard packages such as those implemented in QIIME28 ). For macro-organisms such as mammals, many species have been sequenced for the locus of interest and if not, a closely related species is likely present in the NCBI database (but see also below). Therefore, to analyze DNA sequences from macroorganisms–mammals, amphibians, birds, bryophytes, arthropods, copepods and plants–we used BLAST29 (link) to directly identify the closest DNA sequences in the NCBI database and the likely species of origin. Briefly, we removed from our analyses any DNA sequence observed in less than 10 reads total (summing across all samples), as these likely represent sequencing errors. We then compared each remaining DNA sequence to all sequences deposited in the NCBI nt database using Blastn (excluding uncultured samples) and only considered matches with greater than 90% identity over the entire sequence length. We then retrieved taxonomic data of all best match(es) for each sequence from NCBI. If multiple species matched a single sequence, all species names were assigned to the sequence. We conducted further analyses at the species level for all taxa, using a minimum read count per sample of 10 to determine absence/presence.
While all raw sequences are freely available online (accession number SRP058316), we only describe here the analyses of macroorganism DNA sequences for sake of simplicity (microorganism sequences can be easily analyzed using standard packages such as those implemented in QIIME28 ). For macro-organisms such as mammals, many species have been sequenced for the locus of interest and if not, a closely related species is likely present in the NCBI database (but see also below). Therefore, to analyze DNA sequences from macroorganisms–mammals, amphibians, birds, bryophytes, arthropods, copepods and plants–we used BLAST29 (link) to directly identify the closest DNA sequences in the NCBI database and the likely species of origin. Briefly, we removed from our analyses any DNA sequence observed in less than 10 reads total (summing across all samples), as these likely represent sequencing errors. We then compared each remaining DNA sequence to all sequences deposited in the NCBI nt database using Blastn (excluding uncultured samples) and only considered matches with greater than 90% identity over the entire sequence length. We then retrieved taxonomic data of all best match(es) for each sequence from NCBI. If multiple species matched a single sequence, all species names were assigned to the sequence. We conducted further analyses at the species level for all taxa, using a minimum read count per sample of 10 to determine absence/presence.
Free full text: Click here
