2019
DOI: 10.1016/j.ympev.2019.106562
|View full text |Cite
|
Sign up to set email alerts
|

A demonstration of unsupervised machine learning in species delimitation

Help me understand this report
View preprint versions

Search citation statements

Order By: Relevance

Paper Sections

Select...
2
1
1

Citation Types

11
100
2

Year Published

2019
2019
2024
2024

Publication Types

Select...
6
1

Relationship

1
6

Authors

Journals

citations
Cited by 74 publications
(113 citation statements)
references
References 110 publications
11
100
2
Order By: Relevance
“…Our results confirm that UCEs can provide sufficient variation at shallow timescales in insects to enable species discrimination, adding to previous evidence (Derkarabetian et al, 2019;Harvey et al, 2016;Ješovnik et al, 2017;Ströher et al, 2019;Zarza et al, 2018). Harvey et al (2016) comprehensively compared the utility of sequence capture methods, specifically using UCEs as in our study, and RAD-seq for shallow phylogenies.…”
Section: Concluding Remarks On the Use Of Uces For Species Delimitasupporting
confidence: 85%
“…Our results confirm that UCEs can provide sufficient variation at shallow timescales in insects to enable species discrimination, adding to previous evidence (Derkarabetian et al, 2019;Harvey et al, 2016;Ješovnik et al, 2017;Ströher et al, 2019;Zarza et al, 2018). Harvey et al (2016) comprehensively compared the utility of sequence capture methods, specifically using UCEs as in our study, and RAD-seq for shallow phylogenies.…”
Section: Concluding Remarks On the Use Of Uces For Species Delimitasupporting
confidence: 85%
“…This stems from the many assumptions implicit to MSC, such as random mating, neutral markers, a lack of postspeciation gene flow, and no within‐locus recombination or linkage disequilibrium (Degnan & Rosenberg, 2009). Furthermore, UML methods reduce subjectivity in that they do not rely upon user‐defined models (Derkarabetian, Castillo, Koo, Ovchinnikov, & Hedin, 2019).…”
Section: Methodsmentioning
confidence: 99%
“…Five UML algorithms were also applied (Derkarabetian et al., 2019), to include the following: discriminant analysis of principal components (DAPC: Jombart & Ahmed, 2011); random forest (RF) methods (Liaw & Wiener, 2002) with classical (RF cMDS) and isotonic (RF isoMDS) multidimensional scaling; t‐distributed stochastic neighbor embedding (t‐SNE: van der Maaten & Hinton, 2008); and variational autoencoder (VAE: Kingma & Welling, 2013). For each RF and t‐SNE method, the optimal number of clusters ( K ) was determined by partitioning around medoids (PAM) with gap statistic using k‐means clustering (Kassambara & Mundt, 2019) and via hierarchical clustering (Scrucca, Fop, Murphy, & Raftery, 2016).…”
Section: Methodsmentioning
confidence: 99%
“…Sequence capture methods have great utility at shallow taxonomic levels (Blaimer, LaPolla, Branstetter, Lloyd, & Brady, 2016;Derkarabetian et al, 2019;Hedin, Derkarabetian, Blair, et al, 2018;Smith, Harvey, Faircloth, Glenn, & Brumfield, 2013;Starrett et al, 2017;Tsai et al, 2019;Zarza et al, 2018) including population genomics (Bi et al, 2013), and given the high success in sequencing historical museum specimens, there is great potential for species delimitation. Clear phylogenetic structure and species-level variation are seen across all Vonones samples included in this study, further indicating the potential of incorporating museum specimens into both species-and population-level analyses (Hedin, Derkarabetian, Blair, et al, 2018;Tsai et al, 2019).…”
Section: Sequence Capture Phylogenomics With Historical Specimensmentioning
confidence: 99%
“…In geneious, local nucleotide blast searches were conducted for all assemblies compared to sequences of each gene derived from confamilial/congeneric samples available on GenBank from previously published studies. As the utility of UCE-derived mitochondrial genomes and COI datasets have been demonstrated in previous studies(Derkarabetian, Castillo, Koo, Ovchinnikov, & Hedin, 2019;do Amaral et al, 2015;Hedin, Derkarabetian, Blair, et al, 2018;Zarza et al, 2018), we explored the phylogenetic utility of 28S + 18S rRNA derived from samples included in this study with published data derived from multiple studies…”
mentioning
confidence: 99%