@article{fdi:010094263,
  title = {{M}orphological traits and machine learning for genetic lineage prediction of two reef-building corals},
  author = {{M}itushasi, {G}. and {K}itano, {Y}. {F}. and {O}ury, {N}. and {M}agalon, {H}. and {P}az-{G}arcía, {D}. {A}. and {A}rmstrong, {E}. and {H}ume, {B}. {C}. {C}. and {P}orro, {B}. and {M}oulin, {C}. and {B}oissin, {E}. and {B}ourdin, {G}. and {I}wankow, {G}. and {P}oulain, {J}. and {R}omac, {S}. and {R}eddy, {M}. {M}. and {T}ara {P}acific {C}onsortium {C}oordinators, and {P}lanes, {S}. and {A}llemand, {D}. and {V}oolstra, {C}. {R}. and {F}orcioli, {D}. and {A}gostini, {S}ylvain},
  editor = {},
  language = {{ENG}},
  abstract = {{I}ntegrating multiple lines of evidence that support molecular taxonomy analysis has proven to be a robust method for species delimitation in scleractinian corals. {H}owever, morphology often conflicts with genetic approaches due to high phenotypic plasticity and convergence. {U}nderstanding morphological variation among species is crucial to studying coral distribution, life history, ecology, and evolution. {H}ere, we present an application of {R}andom {F}orest models for coral species identification based on morphological annotation of the corallum and corallites. {W}e show that the integration of molecular and morphological trait analysis can be improved using machine learning. {M}orphological traits were documented for {P}orites and {P}ocillopora coral species that were collected and genotyped through genome-wide, genetical hierarchical clustering, and coalescence analyses for the {T}ara {P}acific {E}xpedition. {W}hile {P}orites only included three tentative species, most {P}ocillopora species were accounted by included specimens from the western {I}ndian {O}cean, tropical {S}outhwestern {P}acific, and southeast {P}olynesia. {T}wo {R}andom {F}orest models per genus were trained on the morphological annotations using the genetic lineage labels. {O}ne model was developed for in-situ image identification and used corallum traits measured from in-situ photographs. {A}nother model for integrative species identification combined corallum and corallite data measured on scanning electron micrographs. {R}andom {F}orest models outperformed traditional dimension reduction methods like {PCA} and {FAMD} followed by k-means and hierarchical clustering by classifying the correct genetic lineage despite morphological clusters overlapping. {T}his machine learning approach is reproducible, cost-effective, and accessible, reducing the need for taxonomic expertise. {I}t can complement molecular and phylogenetic studies and support image identification, highlighting its potential to advance a coral integrative taxonomy workflow.},
  keywords = {{OCEAN} {INDIEN} ; {POLYNESIE} ; {PACIFIQUE} {ILES}},
  booktitle = {},
  journal = {{PL}o{S} {O}ne},
  volume = {20},
  numero = {6},
  pages = {e0326095 [20 p.]},
  year = {2025},
  DOI = {10.1371/journal.pone.0326095},
  URL = {https://www.documentation.ird.fr/hor/fdi:010094263},
}