by Xavier Turon
PopCOmics in the 8th International Barcode of Life Conference
In June 17-20, 2019,
the IBOL Conference was held in Trondheim, Norway. 461 participants
from 61 nations gathered there and discussed about ongoing
developments in barcoding and metabarcoding.
PopCOmics was
represented there by Owen Wangensteen, who delivered a talk about the
novel concept of metaphylogeography developed within PopCOmics… it
enjoyed great acceptance, and Twitter certifies the success of his
presentation!
The abstract
(published in Genome dx.doi.org/10.1139/gen-2019-0097) follows:
From COI
metabarcoding to metaphylogeography: finding the optimal way to get
robust haplotype information from raw HTS data
Owen S. Wangensteen,
Adrià Antich, Creu Palacin, Kim Præbel, and Xavier Turon
Background:
The natural variability of COI makes it the marker of choice in
population genetics and phylogeography studies. The exploitation of
COI metabarcoding datasets for intraspecies patterns of genetic
variation makes it possible to analyse phylogeographic features for
hundreds of species at one time—a field which we name
metaphylogeography. The key to implement this approach is to separate
erroneous sequences from real intra-MOTU variation. This requires a
fine-tuned denoising algorithm to be able to distinguish
low-frequency naturally existing haplotypes from artefacts generated
by sequencing errors.
Results: We calculated the changes in entropy at the different codon positions and the changes in the residual variance of AMOVA models to find optimal filtering parameters in COI metabarcoding datasets. These two independent criteria, coupled with co-occurrence patterns of sequences, can meaningfully guide the cleaning of metabarcoding datasets for metaphylogeography. Using a dataset of community DNA from benthic littoral communities from the Mediterranean and Atlantic seas, we developed a two-step cleaning procedure consisting of a denoising within MOTUs followed by a minimal abundance filtering. Reassuringly, these two independent information-content criteria provided similar optimal parameter values for the denoising and abundance filtering steps, suggesting that the natural haplotype sets can be retrieved. When we compared haplotype networks of two species inferred independently by metaphylogeography and classical phylogeography, the results were also reassuringly similar.
Significance: Our study shows that COI metabarcoding data can be used to infer intra- and interpopulation variability of hundreds of species at one time, providing a new tool with great potential for biogeography, conservation genetics, and invasion genetics.
Results: We calculated the changes in entropy at the different codon positions and the changes in the residual variance of AMOVA models to find optimal filtering parameters in COI metabarcoding datasets. These two independent criteria, coupled with co-occurrence patterns of sequences, can meaningfully guide the cleaning of metabarcoding datasets for metaphylogeography. Using a dataset of community DNA from benthic littoral communities from the Mediterranean and Atlantic seas, we developed a two-step cleaning procedure consisting of a denoising within MOTUs followed by a minimal abundance filtering. Reassuringly, these two independent information-content criteria provided similar optimal parameter values for the denoising and abundance filtering steps, suggesting that the natural haplotype sets can be retrieved. When we compared haplotype networks of two species inferred independently by metaphylogeography and classical phylogeography, the results were also reassuringly similar.
Significance: Our study shows that COI metabarcoding data can be used to infer intra- and interpopulation variability of hundreds of species at one time, providing a new tool with great potential for biogeography, conservation genetics, and invasion genetics.
Group-photo of the
participants at the iBOL 2019 conference
Owen delivering his talk in the Conference (note beard-less Owen!)