Bioinformatics analysis
We conducted the de-multiplexing and de-noising of the samples in the
QIIME2 (version 2019.10) pipeline
(Bolyenet al. 2019) and used the DADA2 method
(Callahan et
al. 2016) to get rid of artefacts and chimeras. Only amplicon sequence
variants (ASVs) that survived the filtering step were kept for
subsequent analyses. We trained a new SILVA V4 Classifier (SSU release
132 515-806) by using QIIME2 tutorials as a reference
(Quastet al. 2012). ASVs were then assigned a taxonomy with the highest
resolution possible (level 7). Following the taxonomic assignment, we
split the analysis into two parts: we kept the original output of the
taxonomy assignment (unfiltered) and then we filtered the chloroplast
and mitochondria assigned reads (filtered). This step was necessary to
evaluate the effect of the clamps on the percentage of reads that were
allocated to the chloroplasts and mitochondria before and after
application of the clamps. The script for our analysis is deposited in
Github (https://github.com/luisvqz/V4_pna_clamps_4_wildlife).
Further analyses were performed in R [version 3.4.4
(2018)] by using
the phyloseq package
(McMurdie & Holmes
2013) in a Linux environment. Plots were generated in R by using the
package ggplot2(Wickham 2016).