3.2 Metabarcoding
We sequenced a total of 9,487,085 reads as part of the DNA metabarcoding
analysis, maintaining 3,196,911 non-chimeric reads after filtering,
denoising, and merging reads for all primer combinations and all samples
(Table 1, Tables S1-S6). The P6 loop of the chloroplast trnL (UAA) intron marker (Figure 3a,b) yielded an average of 56,507 ± 22,886
(SD) non-chimeric processed reads for each of the dissected sections of
the gastrointestinal tract. Consistent with results of the metagenomic
analysis, sequences assigned to the family Erythroxylaceae
(Erythroxylum novagranatense ) predominated in samples from each
section, from the stomach to the C12 portion of the colon (Figure 3a).
Sequences of E. novagranatense represent 97.3% of the 886,078
sequences derived from all surveyed sections of the gastrointestinal
tract. In addition to representatives of Erythroxylaceae, we detected
representatives of Amaryllidaceae (0.4% of the total), Poaceae (0.3%),
Pinaceae (0.3%), and Malvaceae (0.1%), and Fabaceae and Juglandaceae,
which together accounted for 0.05% of the total (Figure 3b).
The internal transcribed spacer 2 (ITS2) marker yielded an average of
19,952 ± 7,762 non-chimeric reads for each section of the
gastrointestinal tract, and Erythroxylaceae accounted for 87.2% of the
279,336 total reads. Sequences from Pinaceae (6.1% of the total),
Amaryllidaceae (5.9%), Urticaceae (0.64%), Poaceae (0.15%) were also
detected, along with traces of Laureaceae, Fabaceae, Ranunculaceae and
Hypericaceae, which together accounted for 0.04% of total reads (Figure
3d).
The third marker that provided information for Streptophyta was domain V
of the 23S plastid rRNA gene which yielded 1,596 ± 671 reads on average
for all sections of the gastrointestinal tract. Again, Erythroxylaceae
predominated, accounting for 98.2% of 22,352 total reads (Figure 3e).
In addition to Erythroxylaceae and Amaryllidaceae (1.5% of total
reads), the next two most abundant taxa were Cyatheaceae and
Brassicaceae, which together accounted for <0.6% of total reads
(Figure 3f).
For Fungi, the internal transcribed spacer (ITS) of the nuclear
ribosomal DNA marker detected a wide variety of Ascomycota and
Basidiomycota orders and families (Figure 4) in the 414,990 read total,
with an average of 29,642 ± 12,846 reads for each section of the
gastrointestinal tract. For Ascomycota, the most abundant families were
Cladosporiacea (28.4% of the total reads for this Phylum),
Pleosporaceae (20.2% of the total), Sacharomycetaceae (16.6% of the
total) and Nectriaceae (14.1% of the total), while the remaining 24
families represent 20.6% of the total, with Phaeococcomycetaceae and
Parmeliaceae (lichen-associated families) accounting for 1.6% of the
total. For the Phylum Basidiocomycota, the orders Agaricales (40.7% of
total reads for this Phylum) and Polyporales (36.5% of the total)
represent the most abundant groups, whereas Psathyrellaceae (18.8% of
the total), Polyporaceae (13.6% of the total), Agaricaceae (12.0% of
the total), and the families Meripilaceae, Fomitopsidaceae and
Hyphodermataceae (together accounting for 16.7% of the total) were the
most abundant groups across all sections of the gastrointestinal tract.
For Metazoans the marker cytochrome c oxidase subunit I identified 41
amplicon sequence variants (ASV), all of which were assigned toPhyllotis ; none were derived from vicuña or guanaco. Nine ASV
remain uncategorized. Finally, the marker cytochrome c oxidase subunit I
specifically developed for arthropods did not detect any ASV for the
group.