Assessing the phylogenetic utility of molecular sequences in Crocodylia

Abstract

Palaeontologists traditionally rely on morphological characters and parsimonyanalyses to reconstruct phylogenetic relationships among organisms. Incontrast, neontologists can use molecular sequences from extant species, apowerful tool for investigating evolutionary relationships. However, not allmolecular sequences are equally valuable for elucidating phylogenetic affinitybetween taxa. While mitochondrial sequences like COX1 and CYTB arewidely used and have a recognised phylogenetic utility, others are includedin phylogenetic analyses without prior assessment. This is relevant for studiesemploying computationally demanding methodologies, such as BayesianInference, where increasing data size does not always improve phylogeneticresolution but significantly increases processing time. Therefore, 33mitochondrial and nuclear nucleotide sequences from 23 extant crocodylianspecies were retrieved from NCBI GenBank to analyse their phylogeneticutility. Each sequence was aligned, tested for substitution saturation(protein-coding genes), and assigned its best nucleotide substitution model.Phylogenetic inference was performed using Bayesian Inference (BI) andMaximum Likelihood (ML), classifying the resulting trees in categories basedon their ability to solve crocodylian relationships. Sequences producingtrees with significant polytomies, mostly partial nuclear sequences, werediscarded, reducing the dataset from approximately 26000 base pairs to17500. The remaining sequences were concatenated and analysed with BIand ML, yielding a highly robust tree congruent with previous molecularphylogenetic hypotheses while optimising computational resources. Theseresults underscore the importance of selecting molecular sequences withphylogenetic utility to improve both the robustness and efficiency of theanalysis of crocodylian evolutionary relationships.