Skip to main content
NCBI home page
Journal of Heredity logoLink to Journal of Heredity
. 2026 Jan 26;117(4):688–698. doi: 10.1093/jhered/esaf106

Effective population size associated with sky-island distribution in the summit rat, Rattus baluensis, a mountain Bornean endemic

Miguel Camacho-Sanchez 1,2,✉, Noor Haliza Hasan 3, Arlo Hinckley 4,5, Fred Tuh Yit Yu 6, Javier Juste 7,8, Jennifer A Leonard 9,✉
PMCID: PMC13326413  PMID: 41582888

Abstract

Tropical mountains are hotspots of biodiversity, and they can harbor species with sky-island distributions on mountaintops. The summit rat, Rattus baluensis, is a Bornean endemic previously known from two populations in the Kinabalu range. Here, we report the discovery of a third, genetically distinct population on Mt. Trusmadi, Sabah, Borneo. Using 27 nuclear introns genotyped in 44 individuals, we show that this population is genetically isolated and more differentiated than those in the Kinabalu range. ABC analyses reveal the three populations likely became isolated in mountain tops during the Holocene. Genetic diversity and effective population size across the three populations correlate with the area of high-elevation habitat on each mountain. Despite relatively large population sizes and well-preserved habitats, the species’ strict association with montane forest and confinement to three mountaintops make it particularly vulnerable to climate change, stochastic events, and localized anthropogenic impacts.

Keywords: genetic diversity, introns, tidyGenR, ABC analysis, Sundaland

Introduction

Tropical regions are renowned for their exceptional biodiversity (Myers et al. 2000). Their mountains, in particular, concentrate a high number of species and endemic lineages (Rahbek et al. 2019). Tropical mountains promote biodiversity through steep environmental gradients, geographic isolation, and ecological specialization (Janzen 1967; Steinbauer et al. 2016; Perrigo et al. 2019). This can lead to sky-island systems—where species restricted to high elevations are isolated on mountaintops (e.g. Gawin et al. 2014; Mastretta-Yanes et al. 2018). Such isolation enhances genetic drift, and, together with climate- and habitat-driven selection, may promote lineage formation (Love et al. 2023).

Sundaland, a geographically diverse region in Southeast Asia, is characterized by a complex topography. The islands of Java and Sumatra are traversed by volcanic mountain chains, many peaks reaching above 3,000 m (e.g. Mt. Kerinci 3,805 m or Mt. Dempo, 3,142 m, in Sumatra), and the Leuser massif, around 3,400 m, occupying northwestern Sumatra. Borneo is dominated by mid-elevation central highlands with some some of the tallest peaks located in the northern region at around 2,500 m (i.e. Mt. Mulu 2,376 m; Mt. Murud, 2,424 m; Mt. Trusmadi, 2,642 m) and Mt. Kinabalu 4,095 m, more isolated in the north. Despite its foundational role in evolutionary theory (Wallace 1858, 1869), Sundaland remains one of the most under-sampled regions for vertebrates (Šmíd 2022). Recent fieldwork and molecular studies have contributed to fill key biodiversity knowledge gaps, known as Linnean (species description) and Wallacean (species distribution) shortfalls (Hortal et al. 2015). They have revealed new mammal species and refined known distributions, highlighting that much of the mammal diversity is associated with mountains—for example, with shrews in Java, Borneo, and Sumatra (Esselstyn et al. 2013; Demos et al. 2016; Hinckley et al. 2022; Nations et al. 2024a), small mammals in Kinabalu National Park (Camacho-Sanchez et al. 2019), and short-tailed gymnures (Hinckley et al. 2024).

In Sabah, northern Borneo, biodiversity surveys date back to the late 19th century (Whitehead 1893), with more recent efforts expanding our understanding of mountain fauna (e.g. Sheldon et al. 2009; Merckx et al. 2015). However, small mammal surveys remain limited and have largely focused on the Kinabalu range (Nor 2001; Wells et al. 2011; Camacho-Sanchez et al. 2019). High-elevation habitats in Borneo host a distinct small mammal community, including many endemics (Camacho-Sanchez et al. 2019). One such species is the summit rat, Rattus baluensis Thomas, 1894, described as endemic to Mt. Kinabalu (4,095 m a.s.l.; Musser 1986). In 2012, a second population was discovered on nearby Mt. Tambuyukon (2,579 m a.s.l.), with both populations tightly associated with montane forest above 2,000 m a.s.l. (Nor 2001; Camacho-Sanchez et al. 2018). Genetic analyses confirmed these populations are currently isolated (Camacho-Sanchez et al. 2018). The summit rat seems to have diverged in the last few hundred thousand years from a widespread lowland species, Rattus tiomanicus. These close relatives exhibit an elevational parapatric distribution (Camacho-Sanchez and Leonard 2020), and are morphologically very similar (Musser 1986). Similarly, Rattus hoogerwerfi and Rattus korinchi inhabit upper mountain ranges in Sumatra (Musser 1986; Camacho-Sanchez and Leonard 2020). These native Sunda rats, along with other species from the Rattus rattus group, seem to have diverged in the last 1.5 My.

Based on its strong association with high-elevation habitat, previous work predicted a third population on Mt. Trusmadi, approximately 58 km south of Kinabalu National Park (Camacho-Sanchez et al. 2018). In this study, we report the discovery of a third population of the summit rat on Mt. Trusmadi. We assessed its genetic diversity and population structure with 20 polymorphic introns. The evolutionary and conservation implications of this finding are discussed.

Material and methods

In July 2016, we surveyed small mammals in the mountain forest along the summit trail of Gunung (Mt.) Trusmadi (also known as Trus Madi) (5.558 N, 116.51E; 2,642 m a.s.l.), Sabah, northern Borneo. This led to the discovery of a third population of summit rats in its montane forest above ~ 2,300 m (Fig. 1; Suppl. Fig. SF1). Summit rats and associated small mammal community were targeted with the Tomahawk and Sherman traps. Half of the Tomahawk traps (n = 20), and half of the Sherman traps (n = 20), were set around the Rest House (~2,050 m a.s.l.), from km 3 to km 2.5 marks along the summit trail, approximately. The remaining 20 Tomahawk and 20 Sherman traps were set in three transects between ~ 2,300 m a.s.l. to the summit. The first transects were set on 6 July 2016, and the last traps were closed 10 days later. The approximate total trapping effort carried out was: 40 traps × 10 nights (400 trap-nights) at lower-mountain forest (2,000 m a.s.l. -2,200 m a.s.l.), 40 traps × 4 nights (160 trap-nights) in the cloud forest between 2,300 m a.s.l. and 2,550 m a.s.l. approximately, and 40 traps × 3 nights (120 trap-nights) on the summit plateau > 2,600 m a.s.l. Up to six individuals per species and elevation site were collected and are currently housed at the Doñana Biological Station mammal collection, Spain (Suppl. Table ST1). Noncollected individuals were processed and released at the trapping site. We sampled muscle or liver tissue for genetic analysis from the collected individuals, whereas for non-collected we sampled ear clips. Recaptured individuals, easily recognized by the ear clips, were released without any further manipulation. The handling protocols adhered to the guidelines of the American Society of Mammalogy (Sikes et al. 2016), and were approved by the CSIC subcommittee of bioethics.

Figure 1.

Figure 1

Open in a new tab

Study area. (A) Map of northern Sabah, Malaysia, Borneo, with its mountain ranges. The area above 2,000 m a.s.l. delimits the likely distribution of the summit rat R. baluensis. The inset on the top-left corner shows its location in Sundaland. (B) Trapping locations along the summit trail on Mt. Trusmadi. (C) Habitat of R. baluensis in the scrublands on the top of Mt. Trusmadi, observed between 2,500 m and the summit in July 2016 (credits: Daniel Hinckley).

Nuclear intron genotyping

We selected a panel of 27 nuclear intron loci to assess the genetic structure and diversity of the new rat population in Trusmadi. This panel contained polymorphic markers for the summit rat that had been previously used successfully to characterize their populations in Kinabalu National Parks mountains (Camacho-Sanchez et al. 2018). Therefore, both the newly generated and previously published data could be integrated to provide a comprehensive assessment of the species’ population genetics.

DNA from around 2 mg of tissue sampled from 44 individuals (Suppl. Table ST1) was extracted using a phenol-chloroform protocol. Its concentration and purity were checked with NanoDrop. The loci were amplified with the primers described in Forcina et al. (2021) and by following a single multiplex (Camacho-Sanchez et al. 2018). Multiplex PCRs took place in 25 μL reactions with 1× Phusion U Green Multiplex PCR Master Mix (Thermo Scientific), and around 250 ng of DNA. All primers had Illumina tails (Supplementary Table ST2). The PCR program consisted of an initial denaturation of 30 s at 98 °C, followed by 30 cycles of denaturation for 10 s at 98 °C, annealing for 30 s with a temperature gradient, starting at 64 °C, 10 cycles Δ-0.5 °C/cycle plus 20 cycles at 59 °C, an extension for 30 s at 72 °C, and a final extension for 10 min at 72 °C. PCR products were purified with Sera-Mag SPRI beads at 1.2× and eluted in TE. Illumina adaptors and sample-specific barcodes were ligated in a second PCR with KAPA HiFi HotStart Ready Mix (Roche) in reactions of 12 μL: 98 °C 30 s, 25 cycles of [98 °C 30s, 56 °C 20s, 72 °C 45 s], and 72 °C 5 min. Libraries were visualized in agarose gels, DNA was quantified using ImageLab (v.6.1) (Bio-Rad), and pooled in approximate equimolarity. Libraries were sequenced on an Illumina MiSeq 300PE at John Hopkins University.

Sequencing reads were demultiplexed by sample based on sample barcodes. Genotypes were called with tidyGenR v0.1.0 (Camacho-Sanchez and Leonard 2026), an R package (R Core Team 2024) for retrieving genotypes from amplicon libraries. Briefly, the standard workflow was followed: raw reads were demultiplexed by locus with demultiplex(), locus-specific truncation thresholds were applied to optimize yield with trunc_amp(), variant calling with variant_call() was carried using an OMEGA_A = 0.01 and BAND_SIZE = 0, frequency filters of minimum allele frequency of 0.1 and allele depth of 10 were applied, and genotypes were called assuming ploidy of 2 with genotype(). A threshold of 10 reads was used to discriminate homozygotes from hemizygotes. Details in the genotyping are available at: https://csmiguel.github.io/tidyGenR_benchmarking.

Population structure and genetic diversity in R. baluensis

To compare the genetic structure and diversity of this new population in Trusmadi within the species diversity and structure, the genotypes of the individuals from Trusmadi were merged with previously published ones from the Kinabalu mountains (Camacho-Sanchez et al. 2018). These could be merged, because both studies used the same panel of loci, into a single tidy data frame using tidyGenR::rename_alleles(). By-locus multiple sequence alignments (MSA) were done with DECIPHER::AlignSeqs(gapOpening = c(−3, −1)) with default parameters but lowering gap opening cost for better alignment of indels in homopolymer regions in loci tmem87a and npr2. All MSAs were inspected visually after DECIPHER::BrowseSeqs() to detect potential non-target variants. The locus fetub was discarded for presenting putative non-target amplicons (also in Camacho-Sanchez et al. 2018), along with six monomorphic loci (apeh17, fancg, irf5, mycbpap, pipox, and ptgs2). A matrix of 20 polymorphic loci in 91 individuals from the three mountains was kept for population genetic analysis (Suppl. Tables ST3 and ST4).

Population structure

Genotypes were re-formatted to a STRUCTURE input. Population structure was assessed with STRUCTURE v2.3.4 (Pritchard et al. 2000), run in 5 cores using structure_threader (Pina-Martins et al. 2017). The runs were executed under the following parameters: clusters K1-K7, 200,000 iterations, 50,000 burnin, four replicates per K, an admixture model, correlated allele frequencies (Falush et al. 2003), and population as priors (Hubisz et al. 2009). Resulting ancestry matrices were aligned with tidyGenR:::align_matrices(). The convergence between the four replicate runs per K was assessed by comparing likelihoods and visualizing aligned ancestry matrices, with pophelper::tabulateQ() and pophelper::plotQ() (Francis 2016), respectively. One unlikely solution for K3 was excluded (Suppl. Fig. SF2), and the remaining replicate runs were averaged. The change of likelihood across K and best K’s was explored with Evanno’s method (Evanno et al. 2005) approach, using pophelper::evannoMethodStructure(). The number of individuals in Trusmadi (44) was around twice compared with Tambuyukon (22) or Kinabalu (25). To evaluate a potential bias from this uneven sampling, we ran two extra STRUCTURE analyses with random subsets from the samples of Trusmadi, 22 individuals in each. The parameters used and downstream plotting was conducted identically to the “complete sampling” run.

Polymorphic genotypes were formatted for GenALEx using tidyGenR::gen_tidy2genalex(). Population differentiation was quantified with an analysis of the molecular variance (AMOVA) from the diploid data (Meirmans 2006) as implemented in GenALEx (Peakall and Smouse 2012). This AMOVA relied on allelic distances between individuals to partition the variance between and within populations. The p-values were estimated after 999 permutations. Genetic structure was further evaluated with pairwise Fst between mountains computed in GenALEx.

Haplotype networks

We built haplotype networks to visualize the diversity and relationships between alleles for all nuclear introns and the potential structure associated with each mountain. The networks for all polymorphic loci were reconstructed using geneHapR (Zhang et al. 2023), which considers indels for the reconstruction of haplotype networks.

Genetic diversity

For each population we computed observed heterozygosity, expected heterozygosity, and the fixation index, calculated per locus and per population in GenALEx. Other sequence-related genetic diversity statistics were computed from the MSA in R: number of polymorphic sites (including indels) with ape::seg.sites() (Paradis et al. 2004), and haplotype diversity and nucleotide diversity (without indels) with pegas (Paradis 2010). A custom R script was used to compute the number of alleles per locus and private alleles per population. Allelic richness was computed with pegas::allelic.richness().

Heterozygosity, elevation, and mountain area

We assessed a possible association of genetic diversity in this island system with “island size,” using as proxies elevation and/or patch size of the elevational range. Per sample multilocus heterozygosity (MLH) was estimated with inbreedR (Stoffel et al. 2016). The area above 2,000 m a.s.l. in each mountain was estimated from a 1-arc­second SRTM digital elevation model in QGIS 3.28 (see map in Fig. 1), as a proxy for the elevational range of R. baluensis. The relation of estimated area of population distribution and MLH was assessed with a linear model in R, in which mountain was added as an interaction (MLH ~ sample_elevation × mountain). The goodness of fit of this full model was contrasted with a reduced model without the interaction via an F-test with anova().

Demographic history

For recent scenarios in which populations have not yet reached monophyly for most loci, studying allele coalescence in relation to their mutation rates, effective population sizes, and population splits can provide insights into demographic histories. We applied the Approximate Bayesian Computation approach implemented in PopABC 1.0 (Lopes et al. 2009) for joint inference (Beaumont et al. 2002) of historical demographic parameters of the summit rat over the last 50 ka. PopABC was run with the same set of 18 loci plus adding the 44 samples from Mt. Trusmadi compared with Camacho-Sanchez et al. (2018). Given the genetic structure described in STRUCTURE analysis (Fig. 2B), and the regression of the upper mountain forest after the LGM (Cannon et al. 2009), we assumed a topology for the population tree in which Mt. Kinabalu and Mt. Tambuyukon are closer ((Kinabalu, Tambuyukon), Trusmadi). We estimated current effective population sizes, ancestral population sizes (NeA1, NeA2), and times of splits (t1 and t2) (Table 1). To simplify inference, migration rates were assumed to be zero. Generation time (1.47 years) and loci-mutation rates (average μ, 4.46 × 10−6 substitutions per generation per locus; SD of μ across loci,1.74 × 10−6) from a similar analysis in Camacho-Sanchez et al. (2018) were used. The units of mutation rates and generation time allow the estimation of times of split in years. A total of 106 simulations were run with the priors in Table 1. Thirteen genetic statistics were computed for each population in the real and simulated datasets. The top 1% of the simulations that yielded most similar genetic statistics compared with the real data were kept in the rejection step. We discarded all simulations in which the effective population size of Mt. Kinabalu exceeded the ancestral NeA1, reflecting our prior expectation that population sizes declined in parallel with the contraction of montane forest since the Last Glacial Maximum. R scripts distributed with PopABC and custom R scripts were used to perform regressions weighted by proximity of simulation, compute, and plot density distribution of the posteriors, and calculate mode and 95% Highest Posterior Density from the posterior distributions.

Figure 2.

Figure 2

Open in a new tab

Population and haplotype relationships and diversity of the summit rat R. baluensis from the three mountains: Mt. Kinabalu, Mt. Tambuyukon, and Mt. Trusmadi. (A) Haplotype networks for the 20 polymorphic loci color-coded by mountain. The size of the circle is proportional to the number of times the allele was observed. Dots represent missing haplotypes. Inset histogram in the bottom right corner depicts the distribution of number of alleles per locus. (B) STRUCTURE ancestries for K2 and K3. (C) Progression of the average likelihood of the data for the four replicate runs from K1 to K7 in STRUCTURE. (D) Individual MHL against elevation, color-coded for each mountain. Each dot represents and individual, with the elevation at which it was caught and its MLH.

Table 1.

Prior and posterior distributions of the parameters used for joint inference with PopABC. Ne, effective population size.

Posterior distribution
Parameter Description Priors Mode 95% HPD
t1 Years to split Ne1-Ne2 Uniform [100,100,00] 1,919 303 to 4,030
t2 Years to split NeA1-Ne3 t1 + uniform [1,020,000] 6,516 3,209 to 11,513
Ne1 Ne in Mt. Kinabalu Uniform [100,100,000] 22,204 14,914 to 28,325
Ne2 Ne in Mt. Tambuyukon Uniform [10,025,000] 2,775 1,929 to 3,701
Ne3 Ne if Mt. Trusmadi Uniform [10,070,000] 4,621 3,489 to 5,707
NeA1 Ancestral Ne of Ne1 + Ne2 Uniform [100,200,000] 43,726 30,476 to 65,393
NeA2 Ancestral Ne before t2 Uniform [100,200,000] 38,552 30,705 to 44,610
μ Average mutation rate 4.46 × 10−6 ... ...
μsd SD of μ across loci 1.74 × 10−6 ... ...
Open in a new tab

Results

Summit rats were most abundant on the summit plateau of Trusmadi, where 30 unique individuals were trapped. The only other species trapped there was the mountain treeshrew Tupaia montana (n = 20). This implied nearly 50% trapping success in the summit area when accounting for recaptures. They were moderately abundant in the cloud forest from 2,300 to around 2,550 m, where we captured 14 individuals, together with mountain spiny rat Crunomys alticola (n = 6), Bornean short-tailed gymnure Hylomys dorsalis (n = 1), Bornean mountain ground squirrel Sundasciurus everetti (n = 3), and mountain treeshrew T. montana (n = 13). However, they were not recorded in the lower-montane mountain forest in Mt. Trusmadi from around 2,000 m to 2,200 m a.s.l. despite our greater trapping effort in this area: 400 trap-nights in lower montane forest as opposed to 280 in the upper montane forest (details in “Materials and Methods”).

The number of reads per sample ranged from 51,500 to 195,000 (mean = 88,300; SD = 28,900). After demultiplexing with locus-specific primers, average reads per sample decreased to 75,000 (SD = 23,500). Reads that passed minimum truncation length and minimal filtering (no N’s and minimal quality), averaged 45,200 (SD = 14,200) per sample. Variant calling and variant filtering with optimal parameters further removed around 1% of the reads to an average of 44,800 reads/sample (SD = 14,100). Diploid genotypes were successfully called for all loci and samples, except for two genotypes (2/1,188 = 0.17% missing calls): tmem87a in BOR1091 and abcg8 in BOR1043 (Suppl. Table ST3). These two genotype calls were dropped to prevent false positives, because three alleles instead of a maximum of two passed the filters. Finally, the genotype dataset contained 91 individual summit rats: 25 from Mt. Kinabalu, 22 from Mt. Tambuyukon, and 44 from Mt. Trusmadi. There were 20 polymorphic loci overall, after removing the six monomorphic ones: apeh17, fancg, irf5, mycbpap, pipox, and ptgs2 (Fig. 2A).

Population structure and genetic diversity

STRUCTURE results showed optimal K values of 2 and 3. K = 3 identified three populations corresponding with the mountain of origin for each sample (Fig. 2B and C). K = 2 identified one ancestral population encompassing all individuals from Kinabalu Park (Mt. Kinabalu + Mt. Tambuyukon), and a second one corresponding to the samples from Mt. Trusmadi. There was no admixture in K2 or K3, with Q-values above 0.95 in all samples. Larger values of K did not point to further population substructure (Suppl. Fig. SF2). Restricting the STRUCTURE runs to even sample sizes between mountains yielded similar results as using the complete dataset (Suppl. Fig. SF3). An AMOVA revealed that 46% (P = 0.001) of the variance in genotypes was explained by populations. Overall Fst was 0.47 (P = 0.001). Pairwise Fst values showed less divergence between Mt. Kinabalu and Mt. Tambuyukon (Fst = 0.12), than those two with Mt. Trusmadi (Fst  trusmadi-tambuyukon = 0.37, Fst  trusmadi-kinabalu = 0.24).

The 44 samples from Mt. Trusmadi had 15 private alleles out of 42 total alleles (36%), whereas the 25 samples from Mt. Kinabalu (4,095 m a.s.l.) had 12 private alleles out from 51 (24%), and the 22 samples from Mt. Tambuyukon (2,579 m a.s.l.) had 4 out of 39 (10%). Mt. Kinabalu and Mt. Tambuyukon shared more alleles than either of them with Mt. Trusmadi (Table 2, Fig. 3; Suppl. Table ST5). Mt. Kinabalu hosted the largest number of alleles (n = 51), followed by Mt. Trusmadi (n = 42) and Mt. Tambuyukon (n = 39) (Table 2). The number of alleles per polymorphic locus ranged from 2 (in 5 loci) to 7 (in tmem87a) (mean = 3.5) (Fig. 2A; Suppl. Table ST6). The average number of SNPs per locus was 2.6 (SD = 1.7), ranging from zero (e.g. p2rx1) to six in rogdi, but the number of polymorphic positions per locus increased to an average of 4.4 (SD = 3.6) when indels were considered. Per-locus diversity statistics showed little differences in observed and expected heterozygosities: Mt. Kinabalu (mean ± SE; He = 0.306 ± 0.052; Ho = 0.289 ± 0.048), Mt. Tambuyukon (He = 0.212 ± 0.048; Ho = 0.218 ± 0.046), and Mt. Trusmadi (He = 0.261 ± 0.046; Ho = 0.256 ± 0.045) (Suppl. Table ST7).

Table 2.

Population genetic estimates for the summit rats from the three mountains and geographic characteristics of the mountains. Number of samples, Ho (observed heterozygosity), He (expected heterozygosity), number of private alleles, allelic richness, Hd (average haplotype diversity across loci with standard deviation), peak height, and mountain area above the 2000 m a.s.l. isoline.

Mt. Kinabalu Mt. Tambuyukon Mt. Trusmadi
No. samples 25 22 44
Ho 0.306 0.212 0.261
He 0.289 0.218 0.256
Private alleles 12 4 15
Allelic richness 51 39 42
Hd (SD) 0.29 (0.22) 0.22 (0.21) 0.26 (0.20)
Peak height (m a.s.l.) 4,095 2,579 2,642
Area > 2000 m (km2)a 113.4 9.8 40.7b
Open in a new tab
a

Used as a proxy for potential distribution of R. baluensis (see Discussion).

b

Area for Trusmadi range, including Gunung Sinsing.

Figure 3.

Figure 3

Open in a new tab

Venn diagram depicting shared and private alleles from all 20 variable nuclear intron loci sequenced from summit rats in the three populations: Mt. Kinabalu, Mt. Tambuyukon, and Mt. Trusmadi.

Heterozygosity and elevation

Observed and expected heterozygosities on Mt. Trusmadi (Ho = 0.26; He = 0.26) were in between those from Mt. Kinabalu (Ho = 0.31; He = 0.29) and Mt. Tambuyukon (Ho = 0.21; He = 0.22). These results point to a positive association between mountain elevation and genetic diversity (Table 3).

Table 3.

Summary of the linear model evaluating the effects of elevation, population, and their interaction on MLH of summit rats. The reference population is that from Mt. Kinabalu.

Estimate SE t value Pr(>|t|)
(Intercept) 0.44 0.061 7.2 <0.001
Elevation −1.3 × 10−4 0.59 × 10−4 −2.3 0.03
Mt. Tambuyukon −0.23 0.083 −2.8 0.006
Mt. Trusmadi −0.31 0.10 −3.1 0.003
elevation × Mt. Tambuyukon 1.7 × 10−4 1.8 × 10−4 0.91 0.37
elevation × Mt. Trusmadi 3.6 × 10−4 1.5 × 10−4 2.4 0.02
Open in a new tab

Bold values indicate statistically significant effects (P < 0.05).

Individual MLH varied between 0 and 0.53 (mean = 0.26, SD = 0.09). There was some evidence of interaction between elevation and mountain as shown by comparing models with/without the interaction (F2,85 4.3; P = 0.05). The full model explained 15% of the variance. On Mt. Kinabalu, elevation had a negative partial effect on MLH (β = −1.3e-04; P = 0.03), whereas on Mt. Trusmadi it was positive (β = 3.6e-04; P = 0.02). Mt. Kinabalu samples had the highest heterozygosity (Fig. 2D; Table 3).

Demographic history

The genetic statistics from our genotypes were coherent with those from simulations (Fig. 4B). The ABC analysis depicted a larger ancestral population which split around 6,516 years ago (95% HPD 3,209 to 11,513) into Mt. Trusmadi and Kinabalu range (NeA1) ones. NeA1 further split at around 1,919 years ago (95% HPD 303 to 4,030) into the current smaller populations of Mt. Kinabalu and Mt. Tambuyukon. The estimated effective population sizes for Mt. Kinabalu (22,204, 95% HPD 14,914 to 28,325), Mt. Tambuyukon (2,775, 95% HPD 1,929 to 3,701) and Mt. Trusmadi (4,621, 95% HPD 3,489 to 5,707), were larger in the mountains with greater available area in the elevational range of the summit rat (Table 2). The certainty of the estimates was highest for the current population sizes and splits but decreased towards the estimation of the ancestral population sizes.

Figure 4.

Figure 4

Open in a new tab

Historical demographic inference of the summit rat with PopABC. (A and C) Prior (dotted lines) and posterior (solid lines) density distribution of each demographic parameter (description in Table 1). (B) Summary statistics from the real data (red dot) and the top 1% simulations are summarized in a PCA. (D) Fixed population branching topology assumed for simulating the data.

Discussion

We describe the genetic structure and diversity of a new summit rat population, and document the first reported outside Kinabalu National Park, in Mt. Trusmadi. Earlier zoological surveys did not detect this species on this mountain. For instance, Harrison (1955) reported stomach contents from all known high-elevation small mammals in the region—such as the mountain giant Sunda rat (Sundamys infraluteus), mountain spiny rat (C. alticola), Bornean mountain ground squirrel (S. everetti), mountain treeshrew (T. montana), and Bornean short-tailed gymnure (H. dorsalis)—but not the summit rat. Similarly, a 1983 expedition by Sheldon and Francis (1985) described the summit ridge as “faunistically depleted,” reporting two species and overlooking other endemics. Probably, their pioneer expedition did not record the summit rat because of the survey methods employed (snares and small metal mammal traps, or “shotguns during the day”).

Our survey revealed a dense population of summit rats and mountain treeshrews on the summit plateau around 2,600 m a.s.l, despite the little available area (Supplementary Fig. SF4). In the cloud forest between 2,300 and 2,550 m, summit rats co-occurred with other montane endemics, including C. alticola, S. everetti, T. montana, H. dorsalis, and possibly Crocidura cf. neglecta (Nor 2001; Camacho-Sanchez et al. 2019; Hinckley et al. 2022). Summit rats were absent from the lower montane forest (2,000 to 2,200 m) in Trusmadi, despite their presence at similar elevations in the Kinabalu range. The summit rat’s absence from other high peaks in Borneo—such as Gunung Alab, Gunung Mulu, and Gunung Murud (Medway 1965; Hanbury-Tenison and Jermy 1979; Wiantoro et al. 2016; Hinckley & Camacho-Sanchez, unpublished; Leonard, unpublished)—supports the idea that only Kinabalu NP and the Trusmadi range (including Gunung Sinsing) host extant populations. Elevational distribution limits may change with the mountain, and population density is patchy. This was evidenced by the higher elevational range of reported summit rats in Trusmadi (> 2,350 m), compared with Tambuyukon (> 2,050 m), and the lower densities in this lower elevational limit compared with larger trapping success in higher elevations in both mountain ranges (Nor 2001; Camacho-Sanchez et al. 2019). In addition, a recent observation of a summit rat at 1,675 m a.s.l. in Kinabalu NP (Quah et al. 2025) suggests local microhabitats—such as areas rich in Nepenthes pitcher plants—may support populations below typical elevational ranges. Kitayama et al. (1993) describes similar vegetation transitions in Trusmadi as those in Kinabalu (Kitayama 1992). Thus, distribution range variation and population density for the summit rat across the different mountains could be related to other factors, such as terrain and vegetation heterogeneity, soils, sun exposure, slope, which could be addressed in future studies.

The summit rat’s distribution exemplifies a sky-island system, where populations are confined to isolated mountaintops. Demographic reconstructions support that the Trusmadi population became isolated approximately 6,500 years ago, whereas the Kinabalu and Tambuyukon populations diverged more recently, around 2,000 years ago. Mt. Trusmadi population shows the highest genetic differentiation, as evidenced by STRUCTURE results, Fst values, and the number of private alleles. The relative isolation of Trusmadi is consistent with the climatic and vegetation history in Sabah, northern Borneo. The Kinabalu massif expands southwestwards to the Crocker Range, but its upper montane forest is disconnected from that in Trusmadi and the mountains in central Borneo by mid elevation areas. The most comprehensive vegetation reconstructions from palynological studies reveal a scenario of oscillating mountain forests extensions and retractions driven by Pleistocene glacial cycles (Cannon et al. 2009). The Last Glacial Maximum, ~ 21 ka, and the following thousands of years marked the latest peak in extension of the upper mountain forest driven by lower temperatures and sea level (Cannon et al. 2009; Hamilton et al. 2024). Under these conditions, we hypothesize one panmictic population of the summit rat across its ancestral range. As temperatures and sea level rose during the Holocene transition, the upper mountain forest retracted upslope. Thus, the summit rat ancestral population would have followed this retraction, isolating the Trusmadi population due to habitat isolation. Secondly, in the late Holocene the nearby populations of Kinabalu and Tambuyukon would have become genetically isolated. As expected in island-like systems, smaller habitat areas are associated with reduced genetic diversity (Frankham 1997; Love et al. 2023). Our data support this: heterozygosity and effective population size correlate with the extent of area above 2,000 m a.s.l. Nonetheless, given that we excluded migration in our ABC analysis, the only forces shaping allele frequency patterns in our models were genetic drift and mutation, so the results should be interpreted cautiously. In the absence of migration, posterior estimates of effective population size (Ne) are likely biased upward, as gene flow can mimic the signal of larger Ne. Similarly, divergence times (Tev) may be biased downward (look more recent), because shared alleles are attributed solely to recent divergence rather than to post-split gene flow.

The pattern of deeper isolation the of Mt. Trusmadi population with regard to Mt. Kinabalu and Mt. Tambuyukon should be evaluated across the mountain small mammal community. In Crocidura cf. neglecta (Hinckley et al. 2022), nuclear and mitochondrial DNA support a greater differentiation of Trusmadi population compared with the ones in Kinabalu and Mount Alab (Crocker Range). However, mtDNA in H. dorsalis suggests a lack of genetic structure between mountains (Hinckley et al. 2024). Broader-ranging species like the mountain treeshrew exhibit gene flow between Mt. Kinabalu and Mt. Tambuyukon ( Parker et al. 2020), but there is no genetic data for its population in Trusmadi, nor for the montane species C. alticola or S. everetti. The varying genetic patterns across the mountain small mammals is also reflected at a broader scale in the varying evolutionary histories of small mammals from Sundaland (Esselstyn et al. 2013; Leonard et al. 2015; Demos et al. 2016; Hinckley et al. 2024; Giarla et al. 2025), when also considering lowland species. This pinpoints that individual species within the same community do not necessarily share similar evolutionary trajectories. When narrowing the comparison down to the native congeneric Rattus, mountain peaks in Sumatran host divergent lineages, R. hoogerwerfi and R. korinchi, 520 to 770 km apart, whereas the Bornean populations we describe, ~ 58 km apart, became isolated only after the LGM (Camacho-Sanchez and Leonard 2020). The shallower divergences of small mammals on Bornean mountains compared with Sumatra also appear to hold for shrews (Demos et al. 2016; Nations et al. 2022). Thus, species across islands separated by sea water or in inland sky-island systems could be experiencing analogous processes, due to geological and climatic drivers with the only difference of scale. In addition, other aspects like species-specific traits, and its derived species-specific niche (e.g. diet, intra, or interspecies interactions, tolerance to abiotic factors), could be relevant in explaining part of the variation in evolutionary patterns, as has been shown in rodents from the nearby island of Sulawesi (Nations et al. 2024b).

Despite relatively high local densities, the summit rat’s isolation, and narrow elevational range make it vulnerable to climate change and stochastic events (Melbourne and Hastings 2008). Warming reduces available habitat, and mountaintops impose hard limits on upward range shifts (Chen et al. 2011; Harris et al. 2012; Camacho-Sanchez et al. 2019). Although Kinabalu NP populations benefit from national park protection, the Mt. Trusmadi population resides in a forestry reserve, offering less formal protection. Nonetheless, steep terrain may provide some natural defense against human disturbance, contrary to extensive degradation of lowland forests across Borneo (Sodhi et al. 2010; Gaveau et al. 2019; Verma et al. 2020).

Conclusion

This study fills a Wallacean shortfall by documenting a previously unknown population of the summit rat on Mt. Trusmadi and analyzing its genetic structure in relation to the two known populations in the Kinabalu range. Mountain habitat discontinuity, modulated by climatic history, seems to have shaped the marked genetic differentiation described between the three ecologically isolated populations. The likely reduction of historical habitat connectivity between mountain peaks probably led to reduced migration and greater drift on the peaks with less available highland area. These results not only enhance our understanding of montane biodiversity in Borneo but also underscore the summit rat’s value as a model for studying sky-island biogeography and recent speciation. Given the species’ vulnerability to climate change and habitat fragmentation, especially outside protected areas, our findings support the need for ongoing attention to the conservation status of this fauna. Future comparative studies across other montane species will be necessary to determine whether similar patterns of isolation and divergence are widespread, and to better understand the evolutionary dynamics of tropical sky islands.

Supplementary Material

SupplementaryFigures_SF1_esaf106
supplementaryfigures_sf1_esaf106.docx (2.1MB, docx)
SuppTables_ST1-ST7_esaf106
supptables_st1-st7_esaf106.xlsx (73.7KB, xlsx)

Acknowledgments

Linus Gonsilou provided logistic support on Mt. Trusmadi. We also want to thank the porters and field assistants Razzak Intang and Cristopher Baidis. The fieldwork was carried out with specific permits of Sabah Parks TTS/IP/100-6/2, the Sabah Forestry Department to do research in Trusmadi JPHTN/TP(FSP) 100-14/18/2/KLT.31. Logistical support was provided by “Laboratorio de Ecología Molecular” (LEM-EBD), and Doñana ICTS-RBD Scientific Collections.

Contributor Information

Miguel Camacho-Sanchez, Conservation & Evolutionary Genetics Group, Estación Biológica de Doñana (EBD-CSIC), Av. Américo Vespucio 26, 41092, Seville, Spain; Institute of Organismic and Molecular Evolution (iomE), Anthropology, Johannes Gutenberg Universität Mainz, Anselm-Franz-von-Bentzel-Weg 7, 55128, Germany.

Noor Haliza Hasan, Institute for Tropical Biology and Conservation, Universiti Malaysia Sabah, (ITBC, UMS), Jalan UMS, 88400 Kota Kinabalu, Sabah, Malaysia.

Arlo Hinckley, Conservation & Evolutionary Genetics Group, Estación Biológica de Doñana (EBD-CSIC), Av. Américo Vespucio 26, 41092, Seville, Spain; Departamento de Zoología, Universidad de Sevilla, Av. Reina Mercedes s/n, 41012, Seville, Spain.

Fred Tuh Yit Yu, Sabah Parks, Block H, Level 1-5, Lot 45 & 46, Signature Office, KK Times Square, 88100, Kota Kinabalu, Sabah, Malaysia.

Javier Juste, Conservation & Evolutionary Genetics Group, Estación Biológica de Doñana (EBD-CSIC), Av. Américo Vespucio 26, 41092, Seville, Spain; Epidemiology and Public Health, CIBERESP, Madrid, Spain.

Jennifer A Leonard, Conservation & Evolutionary Genetics Group, Estación Biológica de Doñana (EBD-CSIC), Av. Américo Vespucio 26, 41092, Seville, Spain.

Author contributions

Miguel Camacho (Conceptualization, Data curation, Formal analysis, Investigation, Methodology, Visualization, Writing—original draft, Writing—review & editing), Noor Haliza Hasan (Project administration, Resources, Writing—review & editing), Arlo Hinckley (Investigation, Writing—review & editing), Fred Tuh Yit Yu (Resources, Writing—review & editing), Javier Juste (Funding acquisition, Project administration, Resources, Writing—review & editing), Jennifer A Leonard (Conceptualization, Funding acquisition, Investigation, Methodology, Project administration, Resources, Supervision, Writing—original draft, Writing—review & editing)

Funding

The project was financed by the Spanish Ministry of Economy and Competition CGL2014-58793-P and the Spanish Ministry of Science and Innovation PID2020-120115GB-100 and PID2023-152577NB-I00.

Data availability

Raw sequencing reads SRA files are deposited in NCBI under the BioProject PRJNA680166. Assembled allele sequences of R. baluensis from Mt. Trusmadi are in NCBI under accessions PQ233918-PQ236201. The code for data analysis and generation of figures has been deposited in GitHub (https://github.com/csmiguel/rattus-baluensis-trusmadi) and in ZENODO (DOI: https://doi.org/10.5281/zenodo.18622241).

References

  1. Beaumont  MA, Zhang  W, Balding  DJ. Approximate Bayesian computation in population genetics. Genetics. 2002;162:2025–2035. 10.1093/genetics/162.4.2025 [DOI] [PMC free article] [PubMed] [Google Scholar]
  2. Camacho-Sanchez  M, Leonard  JA. Mitogenomes reveal multiple colonization of mountains by Rattus in Sundaland. J Hered. 2020;111:392–404. 10.1093/jhered/esaa014 [DOI] [PMC free article] [PubMed] [Google Scholar]
  3. Camacho-Sanchez  M, Quintanilla  I, Hawkins  MTR, Tuh  FYY, Wells  K, Maldonado  JE, Leonard  JA. Interglacial refugia on tropical mountains: novel insights from the summit rat (Rattus baluensis), a Borneo mountain endemic. Divers Distrib. 2018;24:1252–1266. 10.1111/ddi.12761 [DOI] [Google Scholar]
  4. Camacho-Sanchez  M, Hawkins  MTR, Yu  FTY, Maldonado  JE, Leonard  JA. Endemism and diversity of small mammals along two neighboring Bornean mountains. PeerJ. 2019;7:e7858. 10.7717/peerj.7858 [DOI] [PMC free article] [PubMed] [Google Scholar]
  5. Camacho-Sanchez  M, Leonard  J. tidyGenR: Tidy Multilocus Amplicon Genotypes. R package version 0.1.7, 2026. 10.32614/CRAN.package.tidyGenR [DOI] [PMC free article] [PubMed] [Google Scholar]
  6. Cannon  CH, Morley  RJ, Bush  ABG. The current refugial rainforests of Sundaland are unrepresentative of their biogeographic past and highly vulnerable to disturbance. Proc Natl Acad Sci USA. 2009;106:11188–11193. 10.1073/pnas.0809865106 [DOI] [PMC free article] [PubMed] [Google Scholar]
  7. Chen  I-C, Hill  JK, Ohlemuller  R, Roy  DB, Thomas  CD. Rapid range shifts of species associated with high levels of climate warming. Science. 2011;333:1024–1026. 10.1126/science.1206432 [DOI] [PubMed] [Google Scholar]
  8. Demos  TC, Achmadi  AS, Giarla  TC, Handika  H, Maharadatunkamsi  R, K  C, Esselstyn  JA. Local endemism and within-island diversification of shrews illustrate the importance of speciation in building Sundaland mammal diversity. Mol Ecol. 2016;25:5158–5173. 10.1111/mec.13820 [DOI] [PubMed] [Google Scholar]
  9. Esselstyn  JA, Maharadatunkamsi, Achmadi  AS, Siler  CD, Evans  BJ. Carving out turf in a biodiversity hotspot: multiple, previously unrecognized shrew species co-occur on Java Island, Indonesia. Mol Ecol. 2013;22:4972–4987. 10.1111/mec.12450 [DOI] [PubMed] [Google Scholar]
  10. Evanno  G, Regnaut  S, Goudet  J. Detecting the number of clusters of individuals using the software STRUCTURE: a simulation study. Mol Ecol. 2005;14:2611–2620. 10.1111/j.1365-294X.2005.02553.x [DOI] [PubMed] [Google Scholar]
  11. Falush  D, Stephens  M, Pritchard  JK. Inference of population structure using multilocus genotype data: linked loci and correlated allele frequencies. Genetics. 2003;164:1567–1587. 10.1001/jama.1987.03400040069013 [DOI] [PMC free article] [PubMed] [Google Scholar]
  12. Forcina  G, Camacho-Sanchez  M, Tuh  FYY, Moreno  S, Leonard  JA. Markers for genetic change. Heliyon. 2021;7:e05583. 10.1016/j.heliyon.2020.e05583 [DOI] [PMC free article] [PubMed] [Google Scholar]
  13. Francis  RM. POPHELPER: an R package and web app to analyse and visualise population structure. Mol Ecol Resour. 2016;17:27–32. 10.1111/1755-0998.12509 [DOI] [PubMed] [Google Scholar]
  14. Frankham  R. Do island populations have less genetic variation than mainland populations?  Heredity. 1997;78:311–327. 10.1038/sj.hdy.6880980 [DOI] [PubMed] [Google Scholar]
  15. Gaveau  DLA, Locatelli  B, Salim  MA, Yaen  H, Pacheco  P, Sheil  D. Rise and fall of forest loss and industrial plantations in Borneo (2000–2017). Conserv Lett. 2019;12:e12622. 10.1111/conl.12622 [DOI] [Google Scholar]
  16. Gawin  DF, Rahman  MA, Ramji  MFS, Smith  BT, Lim  HC, Moyle  RG, Sheldon  FH. Patterns of avian diversification in Borneo: the case of the endemic mountain black-eye (Chlorocharis emiliae). Auk. 2014;131:86–99. 10.1642/AUK-13-190.1 [DOI] [Google Scholar]
  17. Giarla  TC, Achmadi  AS, Fabre  PH, Handika  H, Chipps  AS, Swanson  MT, Esselstyn  JA. Systematics and historical biogeography of Crunomys and Maxomys (Muridae: Murinae), with the description of a new species from Sulawesi and new genus-level classification. J Mammal. 2025;106:832–858. 10.1093/jmammal/gyaf006 [DOI] [Google Scholar]
  18. Hamilton  R, Amano  N, Bradshaw  CJ, Saltré  F, Patalano  R, Penny  D, Roberts  P. Forest mosaics, not savanna corridors, dominated in Southeast Asia during the last glacial maximum. Proc Natl Acad Sci. 2024;121:e2311280120. 10.1073/pnas.2311280120 [DOI] [PMC free article] [PubMed] [Google Scholar]
  19. Hanbury-Tenison  AR, Jermy  AC. The RGS Expedition to Gunong Mulu, Sarawak 1977-78. The Geographical Journal. 1979;145:175–191. [Google Scholar]
  20. Harris  JBCH, Yong  DL, Sheldon  FH, Boyce  AJ, Eaton  JA, Bernard  H, Biun  A, Langevin  A, Martin  TE, Wei  D. Using diverse data sources to detect elevational range changes of birds on mount Kinabalu, Malaysian Borneo. Raffles Bull Zool. 2012;25:197–247 [Google Scholar]
  21. Hinckley  A, Camacho-Sanchez  M, Ruedi  M, Hawkins  MTR, Mullon  M, Cornellas  A, Yuh F.  TY, Leonard  JA. Evolutionary history of Sundaland shrews (Eulipotyphla: Soricidae: Crocidura) with a focus on Borneo. Zool J Linnean Soc. 2022;194:478–501. 10.1093/zoolinnean/zlab045 [DOI] [Google Scholar]
  22. Hinckley  A, Camacho-Sanchez  M, Chua  MAH, Ruedi  M, Lunde  D, Maldonado  JE, Omar  H, Leonard  JA, Hawkins  MTR. An integrative taxonomic revision of lesser gymnures (Eulipotyphla: Hylomys) reveals five new species and emerging patterns of local endemism in tropical East Asia. Zool J Linnean Soc. 2024;202:zlad177. 10.1093/zoolinnean/zlad177 [DOI] [Google Scholar]
  23. Hortal  J, De Bello  F, Diniz-Filho  JAF, Lewinsohn  TM, Lobo  JM, Ladle  RJ. Seven shortfalls that beset large-scale knowledge of biodiversity. Annu Rev Ecol Evol Syst. 2015;46:523–549. 10.1146/annurev-ecolsys-112414-054400 [DOI] [Google Scholar]
  24. Hubisz  MJ, Falush  D, Stephens  M, Pritchard  JK. Inferring weak population structure with the assistance of sample group information. Mol Ecol Resour. 2009;9:1322–1332. 10.1111/j.1755-0998.2009.02591.x [DOI] [PMC free article] [PubMed] [Google Scholar]
  25. Janzen  DH. Why mountain passes are higher in the tropics. Am Nat. 1967;101:233–249 [Google Scholar]
  26. Kitayama  K. An altitudinal transect study of the vegetation on mount Kinabalu. Borneo Vegetatio. 1992;102:149–171 [Google Scholar]
  27. Kitayama  K, Kulip  J, Nais  J, Biun  A. Vegetation survey on mount Trus Madi, Borneo a prospective new Mountain Park. Mt Res Dev. 1993;13:99–105 [Google Scholar]
  28. Leonard  JA, den  Tex  RJ, Hawkins  MTR, Muñoz-Fuentes  V, Thorington  R, Maldonado  JE. Phylogeography of vertebrates on the Sunda shelf: a multi-species comparison. J Biogeogr. 2015;42:871–879. 10.1111/jbi.12465 [DOI] [Google Scholar]
  29. Lopes  JS, Balding  D, Beaumont  MA. PopABC: a program to infer historical demographic parameters. Bioinformatics. 2009;25:2747–2749. 10.1093/bioinformatics/btp487 [DOI] [PubMed] [Google Scholar]
  30. Love  SJ, Schweitzer  JA, Woolbright  SA, Bailey  JK. Sky Islands are a global tool for predicting the ecological and evolutionary consequences of climate change. Annu Rev Ecol Evol Syst. 2023;54:219–236. 10.1146/annurev-ecolsys-102221-050029 [DOI] [Google Scholar]
  31. Mastretta-Yanes  A, Xue  AT, Moreno-Letelier  A, Jorgensen  TH, Alvarez  N, Piñero  D, Emerson  BC. Long-term in situ persistence of biodiversity in tropical sky islands revealed by landscape genomics. Mol Ecol. 2018;27:432–448. 10.1111/mec.14461 [DOI] [PubMed] [Google Scholar]
  32. Medway  L. Mammals of Borneo: Field keys and an annotated checklist. Journal of the Malayan Branch of the Royal Asiatic Society. 1965;36:1–193. [Google Scholar]
  33. Meirmans  PG. Using the Amova framework to estimate a standardized genetic differentiation measure. Evolution. 2006;60:2399–2402. 10.1111/j.0014-3820.2006.tb01874.x [DOI] [PubMed] [Google Scholar]
  34. Melbourne  BA, Hastings  A. Extinction risk depends strongly on factors contributing to stochasticity. Nature. 2008;454:100–103. 10.1038/nature06922 [DOI] [PubMed] [Google Scholar]
  35. Merckx  VSFT, Hendriks  KP, Beentjes  KK, Mennes  CB, Becking  LE, Peijnenburg  KTCA, Afendy  A, Arumugam  N, de  BH, Biun  A, et al.  Evolution of endemism on a young tropical mountain. Nature. 2015;524:347–350. 10.1038/nature14949 [DOI] [PubMed] [Google Scholar]
  36. Musser  GG. Sundaic Rattus: definitions of Rattus baluensis and Rattus korinchi. Am Mus Novit. 1986;2862:1–24 [Google Scholar]
  37. Myers  N, Mittermeier  RA, Mittermeier  CG, da  FGAB, Kent  J. Biodiversity hotspots for conservation priorities. Nature. 2000;403:853–858. 10.1038/35002501 [DOI] [PubMed] [Google Scholar]
  38. Nations  JA, Giarla  TC, Morni  MA, Dee  JW, Swanson  MT, Hiller  AE, Khan  FAA, Esselstyn  JA. Molecular data from the holotype of the enigmatic Bornean black shrew, Suncus ater Medway, 1965 (Soricidae, Crocidurinae), place it in the genus Palawanosorex. ZooKeys. 2022;2022:17–31. 10.3897/zookeys.1137.94217 [DOI] [PMC free article] [PubMed] [Google Scholar]
  39. Nations  JA, Handika  H, Mursyid  A, Darma Busta  R, Apandi Achmadi  AS, Esselstyn  JA. Three new shrews (Soricidae: Crocidura) from West Sumatra, Indonesia: elevational and morphological divergence in syntopic sister taxa. J Mammal. 2024a;105:372–389 [Google Scholar]
  40. Nations  JA, Kohli  BA, Handika  H, Achmadi  AS, Polito  MJ, Rowe  KC, Esselstyn  JA. The roles of isolation and interspecific interaction in generating the functional diversity of an insular mammal radiation. Oikos. 2024b;2025:1–15. 10.1111/oik.10888 [DOI] [Google Scholar]
  41. Nor  SM. Elevational diversity patterns of small mammals on mount Kinabalu, Sabah. Malaysia Global Ecology and Biogeography. 2001;10:41–62. 10.1046/j.1466-822x.2001.00231.x [DOI] [Google Scholar]
  42. Paradis  E. pegas: an R package for population genetics with an integrated–modular approach. Bioinformatics. 2010;26:419–420. 10.1093/bioinformatics/btp696 [DOI] [PubMed] [Google Scholar]
  43. Paradis  E, Claude  J, Strimmer  K. APE: analyses of phylogenetics and evolution in R language. Bioinformatics. 2004;20:289–290. 10.1093/bioinformatics/btg412 [DOI] [PubMed] [Google Scholar]
  44. Parker  LD, Hawkins  MT, Camacho‐Sanchez M, Campana MG, West‐Roberts JA, Wilbert TR, ..., Maldonado JE. Little genetic structure in a Bornean endemic small mammal across a steep ecological gradient. Molecular Ecology. 2020;29:4074–4090. 10.1111/mec.15626 [DOI] [PubMed] [Google Scholar]
  45. Peakall  R, Smouse  PE. GenALEx 6.5: genetic analysis in excel. Population genetic software for teaching and research. Bioinformatics. 2012;28:2537–2539. 10.1093/bioinformatics/bts460 [DOI] [PMC free article] [PubMed] [Google Scholar]
  46. Perrigo  A, Hoorn  C, Antonelli  A. Why mountains matter for biodiversity. J Biogeogr. 2019;2019:315–325. 10.1111/jbi.13731 [DOI] [Google Scholar]
  47. Pina-Martins  F, Silva  DN, Fino  J, Paulo  OS. Structure_threader: an improved method for automation and parallelization of programs structure, fastStructure and MavericK on multicore CPU systems. Mol Ecol Resour. 2017;17:e268–e274. 10.1111/1755-0998.12702 [DOI] [PubMed] [Google Scholar]
  48. Pritchard  JK, Stephens  M, Donnelly  P. Inference of population structure using multilocus genotype data. Genetics. 2000;155:945–959. 10.1093/genetics/155.2.945 [DOI] [PMC free article] [PubMed] [Google Scholar]
  49. Quah  ESH, Joseph  A, Imbun  PY. Predation on Rattus baluensis (Thomas, 1894) by Trimeresurus sabahi Regenass & Kramer, 1981, in Sabah, Borneo. Herpetology Notes. 2025;18:343–346 [Google Scholar]
  50. R Core Team. R: A Language and Environment for Statistical Computing. R Foundation for Statistical Computing. 2024. https://www.R-project.org/ [Google Scholar]
  51. Rahbek  C, Borregaard  MK, Colwell  RK, Dalsgaard  B, Holt  BG, Morueta-Holme  N, Nogues-Bravo  D, Whittaker  RJ, Fjeldså  J. Humboldt’s enigma: what causes global patterns of mountain biodiversity?  Science. 2019;365:1108–1113. 10.1126/science.aax0149 [DOI] [PubMed] [Google Scholar]
  52. Sheldon  F, Francis  C. The birds and mammals of mount Trusmadi. Sabah SocJ. 1985;VII:77–88 [Google Scholar]
  53. Sheldon  FH, Lim  HC, Nais  J, Lakim  M, Tuuga  A, Malim  P, Majuakim  J, Lo  A, Schilthuizen  M, Hosner  PA. Observations on the ecology, distribution, and biogeography of forest birds in Sabah. Malaysia Raffles Bull Zool. 2009;57:577–586 [Google Scholar]
  54. Sikes  RS, the Animal Care and Use Committee of the American Society of Mammalogists . Guidelines of the American Society of Mammalogists for the use of wild mammals in research and education. J Mammal. 2016;97:663–688. 10.1093/jmammal/gyw078 [DOI] [PMC free article] [PubMed] [Google Scholar]
  55. Šmíd  J. Geographic and taxonomic biases in the vertebrate tree of life. J Biogeogr. 2022;49:2120–2129. 10.1111/jbi.14491 [DOI] [Google Scholar]
  56. Sodhi  NS, Posa  MRC, Lee  TM, Bickford  D, Koh  LP, Brook  BW. The state and conservation of Southeast Asian biodiversity. Biodivers Conserv. 2010;19:317–328. 10.1007/s10531-009-9607-5 [DOI] [Google Scholar]
  57. Steinbauer  MJ, Field  R, Grytnes  JA, Trigas  P, Ah-Peng  C, Attorre  F, Birks  HJB, Borges  PAV, Cardoso  P, Chou  CH, et al.  Topography-driven isolation, speciation and a global increase of endemism with elevation. Glob Ecol Biogeogr. 2016;25:1097–1107. 10.1111/geb.12469 [DOI] [Google Scholar]
  58. Stoffel  MA, Esser  M, Kardos  M, Humble  E, Nichols  H, David  P, Hoffman  JI. inbreedR: an R package for the analysis of inbreeding based on genetic markers. Methods Ecol Evol. 2016;7:1331–1339. 10.1111/2041-210X.12588 [DOI] [Google Scholar]
  59. Verma  M, Symes  WS, Watson  JEM, Jones  KR, Allan  JR, Venter  O, Rheindt  FE, Edwards  DP, Carrasco  LR. Severe human pressures in the Sundaland biodiversity hotspot. ConservSciPract. 2020;2:e169. 10.1111/csp2.169 [DOI] [Google Scholar]
  60. Wallace  AR. On the tendency of varieties to depart indefinitely from the original type. Journal of the Proceedings of the Linnean Society of London, Zoology. 1858;3:53–62. [Google Scholar]
  61. Wallace  AR. The Malay archipelago: the land of the orang-utan, and the bird of paradise. A narrative of travel, with studies of man and nature. En Malay Archipelago. London: Macmillan & Co; 1869. [Google Scholar]
  62. Wells  K, Lakim  MB, Beaucournu  J-CC. Host specificity and niche partitioning in flea-small mammal networks in Bornean rainforests. Med Vet Entomol. 2011;25:311–319. 10.1111/j.1365-2915.2010.00940.x [DOI] [PubMed] [Google Scholar]
  63. Whitehead  J. Exploration of mount kina Balu. North Borneo: Gurney and Jackson; 1893. [Google Scholar]
  64. Wiantoro  S, Lit  E, Sidq  MF, Salmizar  N, Sait  IAbdullah M. Notes on Field Survey and New Distributional Record of Small Mammals in Mount Murud, Sarawak, Malaysia. Journal of Tropical Biology & Conservation. 2016;5:71–79. [Google Scholar]
  65. Zhang  R, Jia  G, Diao  X. geneHapR: an R package for gene haplotypic statistics and visualization. BMC Bioinform. 2023;24:199. 10.1186/s12859-023-05318-9 [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

SupplementaryFigures_SF1_esaf106
supplementaryfigures_sf1_esaf106.docx (2.1MB, docx)
SuppTables_ST1-ST7_esaf106
supptables_st1-st7_esaf106.xlsx (73.7KB, xlsx)

Data Availability Statement

Raw sequencing reads SRA files are deposited in NCBI under the BioProject PRJNA680166. Assembled allele sequences of R. baluensis from Mt. Trusmadi are in NCBI under accessions PQ233918-PQ236201. The code for data analysis and generation of figures has been deposited in GitHub (https://github.com/csmiguel/rattus-baluensis-trusmadi) and in ZENODO (DOI: https://doi.org/10.5281/zenodo.18622241).

Articles from Journal of Heredity are provided here courtesy of Oxford University Press

RESOURCES