Metabarcoding proof
Finally, based on taxonomic amplification and resolution capacities
revealed in the “in silico” PCRs, we selected the MiniB18S_81
mini-barcode for assessment in a real metabarcoding analysis. The
samples tested were faeces samples collected (in Spain) from 12 bird
flocks of 5 different steppe bird species; 1 flock of farmed red-legged
partridges (Alectoris rufa ), 2 flocks of wild red-legged
partridges (Alectoris rufa ), 4 flocks of pin-tailed sandgrouse
(Pterocles alchata ), 1 flock of black-bellied sandgrouse
(Pterocles orientalis ), 2 flocks of great bustards (Otis
tarda ) and 2 flocks of little bustards (Tetrax tetrax ).
Whenever possible, fresh faeces from 20 individuals in each flock were
collected. Faeces were collected in the roosts without trapping any
birds. DNA extractions were performed using the QIAamp® DNA Stool Mini
Kit from Quiagen (Ref. 51504). From 15 to 20 DNA samples were obtained
per flock (DNA could not be extracted from every sample). The
concentration (μg/ml) and quality of each DNA sample were measured using
the NanoVue Plus spectrophotometer (GE Healthcare). Samples from each
flock were mixed according to DNA quantity and quality in 3 pools of 5-7
samples, one of high, one of medium and one of low quantity.
Subsequently, DNA was PCR amplified from extracts using each of the
designed primers (miniB18S_81F and miniB18S_81R) (Table 1) at the
Analytical Services (SGIker) of the University of the Basque Country,
UPV/EHU. To retrieve enough material for the sequencing reaction, three
PCR reactions were run per sample pool. Samples were purified and a
second reaction was performed to index each amplified product and attach
Illumina adaptors using the Illumina Nextera v2 kit. We used the same
Illumina adaptor for the three pools prepared for each flock.
Amplification (20 µl) was carried out using 1 µl of DNA in a reaction
mix containing 1.25 µl of MgCl (25 mM), 2 µl of dNTPs (20 mM), 0.25 of
10X Buffer B, 0.3 µl of each forward (20 pM) and reverse (20 pM) primer,
1 µl of BSA (10mg/ml) and 0.25 µl of GoTaq® Flexi DNA Polymerase from
Promega (ref. M829; 5 U/µl). The thermocycle conditions were: 96°C for 1
minute and 35 cycles of 94°C for 30 seconds, 57°C for 30 seconds, 72°C
for 1 minute. Once amplified, PCR outputs were sequenced in an Illumina
MiSeq NGS platform (sequencing of 2x150 bp paired-end reads) with the
MiSeq Reagent Kit v2, following the manufacturer’s instructions.
Bioinformatic sequencing output tests were conducted using the Cutadapt
(Martin, 2011) and Usearch (Edgar, 2010) software packages. First, the
adapters were cut and the forward and reverse sequences combined. The
sequences were then cleaned and OUT tables built based on 97% OTUs. We
used the join.fastq function of Usearch instead of merging
forward and reverse sequences to avoid losing sequences of amplicons
longer than 300 bp. Subsequently, the sintax function of Usearch
was used to predict the taxonomy of the OTUs obtained with acutoff of 0.8 and a combined 18S and 16S dataset. The
reference dataset was built combining the Silva 18S v123 dataset (Quast
et al., 2013) with the RDP 16S v16 training set (Cole et al., 2014),
both provided by the Usearch platform. We opted for the RDP 16S v16
training dataset instead of a complete Silva, Greengenes or RDP 16S
dataset because of the high level of mistaken sequences reported for
these datasets (Edgar, 2018).
All graphs were constructed using the ggplot2 package (Wickham
2016) in R v3.6.1 (R Core Team 2019).