A combined approach to identify isolated theropod teeth from the Cenomanian Kem Kem Group of Morocco: cladistic, discriminant, and machine learning analyses

ABSTRACT

The Kem Kem Group of Southeastern Morocco, North Africa, is well known for theropod remains, especially isolated teeth. Here, a collection of isolated theropod teeth is assessed for diversity using a combination of linear discriminant, phylogenetic, and machine learning analyses for the first time. The results confirm earlier studies on Kem Kem theropod diversity, with teeth referred to Abelisauridae, Spinosaurinae, and Carcharodontosauridae. A single tooth is ascribed to a non-abelisauroid ceratosaur or a megaraptoran and may represent the enigmatic averostran Deltadromeus. Spinosaurine teeth are clearly differentiated by all three methodologies, whereas abelisaurid and carcharodontosaurid teeth could only be distinguished by the machine learning and phylogenetic analyses. This study shows that a combination of independent methods is most effective at providing strong evidence on theropod dental diversity in a particular assemblage, and that cladistic and machine learning analyses are the most reliable approaches to identify isolated dinosaur teeth. The methodology used here is likely to yield results in other dinosaur assemblages where isolated teeth are more abundant than body fossils.

DISCLOSURE STATEMENT

No potential conflict of interest was reported by the author(s).

DATA AVAILABILITY STATEMENT

The dentition and crown-based data matrices (NEXUS and TNT files) used in the cladistic analyses are available on Morphobank: http://morphobank.org/permalink/?P4725

ACKNOWLEDGMENTS

We are grateful to E. Buffetaut (CNRS) and E. Malafaia (Uni. Lisboa) for sharing information and papers useful for this study. We are indebted to Charles Underwood for his generous aid providing a stratigraphic figure of the Kem Kem Group around Taouz. We acknowledge the use of the Willi Hennig Society edition of TNT for the cladistic analysis and Phylopic for the theropod silhouettes, and thank Scott Hartman, T. Dixon, J. F. Designs, M. Kevil, V.-V. Sinkkonen, J. Garza, and T. Tischler for sharing their original artworks on this website. We additionally thank JVP editor G. Bever, phylogenetic editor P. L. Godoy, and two anonymous reviewers for their valuable comments, corrections, and suggestions, which significantly improved the final version of this manuscript. C.H. was supported by the Consejo Nacional de Investigaciones Científicas y Técnicas (CONICET) and Agencia Nacional de Promoción Científica y Tecnológica, Argentina (Beca Post-doctoral CONICET Legajo 181417). F.M.H. was funded by a scholarship supporting faculty-specific gender equality targets at Friedrich-Alexander University Erlangen-Nürnberg (FAU), and by the Dr Elizabeth Nicholls Postdoctoral Fellowship of the Royal Tyrrell Museum Cooperating Society (RTMCS). T.H.T. was funded by a studentship from the Natural Environment Research Council GW4+ Doctoral Training Partnership (NERC GW4+). C.H. dedicates this work to M.G. and M.N. Martin.

AUTHOR CONTRIBUTIONS

FMH, JWFR, THT, OWMR, RRM, and JVW designed the project. CH, THT, FMH, KHWS, JVW, and RRM gathered the data. CH conducted the phylogenetic and discriminant analyses. SW conducted the machine learning analysis. CH and SW analyzed the data. CH, THT, and FMH drafted the manuscript. CH, THT, SW, FMH, KHWS, OWMR, and JWFR edited the manuscript.

SUPPLEMENTARY FILES

Supplementary File 1.docx: dentition-based characters, dentition- and crown-based datamatrices, and results of the cladistic analyses. [Theropod_teeth_Kem_Kem_Group_JVP_sup_inf_new.docx]

Supplementary File 2.xlsx: dataset on crown-based measurements and results of the discriminant and machine learning analyses. [Crown measurement dataset Kem Kem theropods new.xlsx]

Supplementary File 3.xlsx: dentition and crown-based data matrices, results of the cladistic analyses, and phylopic licenses. [Dentition-based datamatrix Kem Kem theropods Jul 2023.xlsx]

Supplementary File 4.Rmd: machine learning models used in the machine learning analyses. [KemKemMachineLearingModels.Rmd]

Supplementary File 5.dat: Personal dataset of crown-based measurements used in Past 3.11 and considering the absence of denticles as missing data. [Morpho Kem Kem teeth My dataset unserr new.dat]

Supplementary File 6.dat: Personal dataset of crown-based measurements used in Past 3.11 and using an arbitrary value of 100 denticles per 5 mm for non-serrated carinae. [Morpho Kem Kem teeth My dataset.dat]

Supplementary File 7.dat: Whole dataset of crown-based measurements used in Past 3.11 and considering the absence of denticles as missing data. [Morpho Kem Kem teeth whole dataset unserr.dat]

Supplementary File 8.dat: Whole dataset of crown-based measurements used in Past 3.11 and using an arbitrary value of 100 denticles per 5 mm for non-serrated carinae. [Morpho Kem Kem teeth whole dataset.dat]

Supplementary File 9.dat: Personal dataset of crown-based measurements restricted to large sized crowns (CH > 20 mm) used in Past 3.11 and considering the absence of denticles as missing data. [Morpho large Kem Kem teeth My dataset unserr new.dat]

Supplementary File 10.dat: Personal dataset of crown-based measurements restricted to large sized crowns (CH > 20 mm) used in Past 3.11 and using an arbitrary value of 100 denticles per 5 mm for non-serrated carinae. [Morpho large Kem Kem teeth My dataset.dat]

Supplementary File 11.dat: Whole dataset of crown-based measurements restricted to large sized crowns (CH > 20 mm) used in Past 3.11 and considering the absence of denticles as missing data. [Morpho large Kem Kem teeth whole dataset unserr new.dat]

Supplementary File 12.dat: Whole dataset of crown-based measurements restricted to large sized crowns (CH > 20 mm) used in Past 3.11 and using an arbitrary value of 100 denticles per 5 mm for non-serrated carinae. [Morpho large Kem Kem teeth whole dataset.dat]

Supplementary File 13.tree: complete set of 10,000 most parsimonious trees obtained in the parsimony analysis performed on the crown-based datamatrix with all morphotypes included and no constrained tree. [MPTs_crown_all_morphotypes.tre]

Supplementary File 14.tree: complete set of 12 most parsimonious trees obtained in the parsimony analysis performed on the dentition-based datamatrix with all morphotypes included and using a constrained tree. [MPTs_dentition_constrained_all_morphotypes.tre]

Supplementary File 15.tree: complete set of two most parsimonious trees obtained in the parsimony analysis performed on the dentition-based datamatrix using Morphotype I as a separate OTU and a constrained tree. [MPTs_dentition_constrained_I.tre]

Supplementary File 16.tree: complete set of three most parsimonious trees obtained in the parsimony analysis performed on the dentition-based datamatrix using Morphotype Ia as a separate OTU and a constrained tree. [MPTs_dentition_constrained_Ia.tre]

Supplementary File 17.tree: complete set of two most parsimonious trees obtained in the parsimony analysis performed on the dentition-based datamatrix using Morphotype Ib as a separate OTU and a constrained tree. [MPTs_dentition_constrained_Ib.tre]

Supplementary File 18.tree: complete set of two most parsimonious trees obtained in the parsimony analysis performed on the dentition-based datamatrix using Morphotype II as a separate OTU and a constrained tree. [MPTs_dentition_constrained_II.tre]

Supplementary File 19.tree: complete set of a single most parsimonious trees obtained in the parsimony analysis performed on the dentition-based datamatrix using Morphotype III as a separate OTU and a constrained tree. [MPTs_dentition_constrained_III.tre]

Supplementary File 20.tree: complete set of a single most parsimonious trees obtained in the parsimony analysis performed on the dentition-based datamatrix using Morphotype IIIa as a separate OTU and a constrained tree. [MPTs_dentition_constrained_IIIa.tre]

Supplementary File 21.tree: complete set of 11 most parsimonious trees obtained in the parsimony analysis performed on the dentition-based datamatrix using Morphotype IIIb as a separate OTU and a constrained tree. [MPTs_dentition_constrained_IIIb.tre]

Supplementary File 22.tree: complete set of three most parsimonious trees obtained in the parsimony analysis performed on the dentition-based datamatrix using Morphotype IV as a separate OTU and a constrained tree. [MPTs_dentition_constrained_IV.tre]

Supplementary File 23.tree: complete set of 10,000 most parsimonious trees obtained in the parsimony analysis performed on the dentition-based datamatrix with all morphotypes included and no constrained tree. [MPTs_dentition_unconstrained_all_morphotypes.tre]

Supplementary File 24.nex: NEXUS file of the crown-based datamatrix. [Theropod Kem Kem Crown datamatrix Jun 2023.nex]

Supplementary File 25.nex: NEXUS file of the dentition-based datamatrix. [Theropod Kem Kem Dentition datamatrix Jun 2023.nex]

Supplementary File 26.tnt: TNT file of the crown-based datamatrix. [Theropod Kem Kem Crown datamatrix Jun 2023.tnt]

Supplementary File 27.tnt: TNT file of the dentition-based datamatrix. [Theropod Kem Kem Dentition datamatrix Jun 2023.tnt]