Identifying a new phylogeographic population of Blyth’s Tragopan (Tragopan blythii) through multi-locus analyses

Published: 22-03-2021| Version 1 | DOI: 10.17632/jcs246skcn.1
De Chen


Geographically and morphologically distinct populations within a species are frequently the focus of conservation efforts, especially when the populations are evolutionarily significant units. Blyth’s Tragopan (Tragopan blythii) is a globally-threatened species confined to South and Southeast Asia. During our field surveys in western Myanmar, we discovered a distinct group of individuals that differed in their appearance relative to all other populations. We further examined the differences in their DNA sequence using three nuclear introns and three mitochondrial genes through phylogenetic analytical methods. Our results showed the population from Mount Kennedy, Chin Hills formed reciprocal monophyletic groups with the nominate subspecies from Mount Saramati, Sagaing Division. Species delimitation analyses further confirmed this differentiation. Geographical isolation by the intervening lowlands found between high elevation habitats may have been the main cause of their differentiation. Hence, we propose that the Mount Kennedy population be viewed as a distinct evolutionarily significant unit and be given special priority for conservation.


Steps to reproduce

We amplified three mtDNA genes, cytochrome oxidase subunit 1 (CO1), cytochrome b (CYTB), NADH dehydrogenase subunit 2 (ND2), three nuclear introns, clathrin heavy chain-like 1 (CLTCL1) intron 7, fibrinogen beta chain (FGB) intron 5, and ovalbumin (SERPINB14) intron 3. We combined the six partitions into three combined datasets in the phylogenetic analyses, mtDNA, nuclear and combined sequences. Maximum likelihood (ML) analyses were performed in RAxML v8.2.12 (Stamatakis 2014) using the GTRGAMMA model in each partition with the ‘-f a’ option that generates the optimal tree and conducts 100 rapid bootstrap searches. Then, Bayesian inference (BI) was performed in BEAST v1.10.4 (Suchard et al. 2018) using the best-fitting nucleotide substitution model in each partition obtained from jModelTest v2.1.7 (Darriba et al. 2012). BI were run for 10,000,000 generations and the chains were sampled every 1,000 generations. The posterior probabilities were assessed using Tracer v1.7.1 (Rambaut et al. 2018) to ensure that the effective sample size was larger than 200. The first 25% of the total number of generated trees were discarded as burn-in, while the remaining trees were used to calculate the consensus tree using TreeAnnotator v1.10.4 (Suchard et al. 2018).