﻿Novel genomic resources contribute to the systematics of threatened arboreal deer mice of the genus Habromys Hooper & Musser, 1964 (Cricetidae, Neotominae) within a neotomine–peromyscine phylogeny

﻿Abstract The Crested-tailed deer mouse, Habromyslophurus, is one of seven arboreal species within the genus Habromys. Species of this genus are monotypic, relatively rare, and occur in low densities. Their geographical distribution is highly fragmented due to being restricted to montane cloud forest in Mesoamerica and they are of conservation concern. All Habromys species are endemic to Mexico, except H.lophurus, which is also distributed in Guatemala and El Salvador. In this study, we obtained and characterized the first mitogenome and several thousand nuclear ultraconserved elements (UCEs) of H.lophurus to determine its phylogenetic position within neotomine–peromyscine mice. Its mitogenome sequence (16,509 bp) is only the second complete mitogenome obtained for this poorly known genus. We also obtained the first nuclear genomic data for H.lophurus, including 3,654 UCE loci, as well as a partial mitogenome of H.simulatus (6,349 bp), and 2,186 UCE for the outgroup Holochilussciureus. Phylogenetic analyses that included our newly generated genomic data coupled with previously published data from other neotomine–peromyscine mice confirm the placement of H.lophurus, H.simulatus, and H.ixtlani within a highly supported clade. The Habromys clade was nested within a clade that also contains members of the genus Peromyscus and provides further support for the hypothesis of the paraphyly of Peromyscus. These genomic resources will contribute to future phylogenomic studies that aim to further elucidate the evolutionary history of this rare and critically endangered genus of rodents.


Introduction
The genus Habromys Hooper & Musser, 1964, considered a relict, comprises seven monotypic species (H.simulatus (Osgood, 1904), H. chinanteco (Robertson & Musser, 1976) Vidales, 2002, H. schmidlyi Romo-Vázquez, León-Paniagua & Sánchez, 2005, H. lophurus (Osgood, 1904), H. lepturus (Merriam, 1898), and H. ixtlani (Goodwin, 1964)) which inhabit montane regions of Mesoamerica, ranging from central-southern Mexico to north-eastern El Salvador (Hooper 1968;Carleton et al. 2002;León-Paniagua et al. 2007;Rogers et al. 2007).Species of Habromys are arboreal and restricted to montane cloud forest and moist-oak forest.The montane cloud forest, a relict ecosystem, is one of the most highly endangered and fragmented habitats mainly because of land conversion for grazing and agriculture (Toledo-Aceves et al. 2011).The association between species of the genus Habromys and this ecosystem makes their geographical distribution highly discontinuous and susceptible to climate change and anthropogenic activities (Carleton et al. 2002;Contreras-Medina et al. 2022).These species are elusive and rare, with small population sizes, and are poorly represented in scientific collections (Romo-Vázquez et al. 2005;León-Paniagua et al. 2007;Castañeda-Rico et al. 2011).In addition, each species is currently found in only one or a few localities, resulting in micro-endemic distributions (León-Paniagua et al. 2007;Rogers et al. 2007;Castañeda-Rico et al. 2011).The characteristics mentioned above increase the risk of extinction for all species within this genus.Members of the genus Habromys are considered characteristic elements of the Neotropical montane cloud forest and contribute to the recognition of Mesoamerica as a biodiversity hotspot (Contreras-Medina et al. 2022).Despite the biological importance of this genus, little is known about its evolutionary history.León-Paniagua et al. (2007) and Rogers et al. (2007), using mitochondrial genes (ND3, ND4L, partial ND4, and Cyt b, respectively), conducted the only two molecular phylogenetic studies of the genus to date.
The crested-tailed deer mouse, Habromys lophurus, is distributed in Nuclear Central America, from Chiapas, Mexico to Guatemala and El Salvador (León-Paniagua et al. 2007;Rogers et al. 2007).It is possible that this species also occurs in small patches in western Honduras (Carleton et al. 2002;Contreras-Medina et al. 2022) but, to date, no specimens have been identified or captured in that region.Robertson and Musser (1976) and Carleton et al. (2002) found that the body size in populations from Chiapas is smaller than those in Guatemala and El Salvador, but they considered that this distinction did not warrant taxonomic recognition.However, León-Paniagua et al. (2007), using three mitochondrial genes, suggested that both H. lophurus and H. simulatus consist of two geographically isolated and restricted species; i.e. H. lophurus from the highlands of Chiapas, Mexico, and northern Guatemala, and H. sp.nov. 2 from Sierra Las Minas, eastern Guatemala to El Salvador, and H. simulatus from central Sierra Madre Oriental, Mexico and H. sp.nov. 1 from the Sierra Mazateca, Oaxaca, Mexico.In addition, Castañeda-Rico et al. (2011), using microsatellites, found that H. simulatus included two distinct evolutionary units.The objective of this study was to generate additional mitochondrial data representing the full mitochondrial genome, as well as the first nuclear genomic data ever generated for the species-i.e.thousands of ultraconserved element loci-to place H. lophurus within a neotomine-peromyscine phylogeny and to provide genomic resources for future phylogenomic studies with a denser taxon sampling.

Specimen sampling and lab work
We used a tissue sample of Habromys lophurus associated with a male museum voucher specimen (Fig. 1), both of which are accessioned at the Smithsonian Institution's National Museum of Natural History with catalog number USNM 569392, to sequence the first publicly available mitogenome and several thousand nuclear ultraconserved elements (UCEs) of this species.The specimen was collected in 2005 in Guatemala, Huehuetenango, 5 km southwest of San Mateo Ixtatán (15°48′36.36″N,91°30′19.44″W).
DNA was extracted from the tissue sample (ca 1.5 mm 2 of frozen internal organ) using the DNeasy Blood and Tissue Kit (QIAGEN, Inc., Valencia, CA, USA), following the manufacturer's protocol, and quantified using a Qubit (Thermo Fisher, Waltham, MA, USA) fluorometer with a 1× dsDNA HS assay kit.DNA was sheared to an average length of 250 base pairs (bp) using a Bioruptor Pico sonicator (Diagenode Inc., Denville, NJ, USA) with a pulse of 30 s on, 30 s off for 90 cycles.A dual-indexed library with TruSeq-style indices (Meyer and Kircher 2010), amplified with Kapa HiFi HotStart Ready Mix (Roche Sequencing), was prepared using the Kapa HyperPrep kit (Roche Sequencing, Indianapolis, IN, USA) with ½ reactions following the manufacturer's protocol.The myBaits Mito kit for the house mouse (Mus musculus Linnaeus, 1758) and the myBaits UCE Tetrapods 5Kv1 kit, both produced by Daicel Arbor Biosciences, were used for target-enrichment to capture mitogenomes and UCEs, respectively.The post-enrichment libraries were amplified with 14 cycles of PCR using universal Illumina primers and Kapa HiFi Hotstart Ready Mix, and sequenced on a NovaSeq 6000 (Illumina, Inc., San Diego, CA, US) at the Oklahoma Medical Research Foundation, Oklahoma City (combined with samples from other projects).We also reanalyzed raw data published by Castañeda-Rico et al. (2020, 2022a), under BioProjects PRJNA606805 and PRJNA818347, to obtain a mitogenome of Habromys simulatus and UCEs of Holochilus sciureus Wagner, 1842.

Bioinformatics and phylogenetic analyses
For mitogenomes, read quality analysis of the FASTQ format files was conducted using FastQC v0.11.5 (Andrews 2010).Low-quality reads and adapter contamination were removed using TrimGalore v0.6.5 (https://github.com/Felix-Krueger/TrimGalore).Prinseq-lite v0.20.4 (Schmieder and Edwards 2011) was used to remove exact duplicates (-derep1,4).Cleaned reads were mapped to a reference (Habromys ixtlani GenBank KY707304) using the Geneious algorithm in Geneious Prime 2021.2.2 (https://www.geneious.com)with default parameters (Medium-Low sensitivity, Maximum mismatches = 20%, Maximum gaps = 10%).Consensus sequences were generated with Geneious, using 5X as the lowest coverage to call a base.The MAFFT v7.45 plug-in (Katoh and Standley 2013) in Geneious was used to align the consensus sequences and references.We transferred annotations from the reference mitogenome and translated all protein-coding genes (PCGs) to rule out the presence of nuclear copies of mitochondrial genes (NUMTs) and to check for frame shifts or stop codons.
To determine the phylogenetic position of H. lophurus, a Bayesian Inference (BI) analysis was performed using MrBayes v3.2.6 (Ronquist et al. 2012) under the best model and partition scheme (by gene and codon position -6 partitions) selected by PartitionFinder v2.1.1 (Lanfear et al. 2016) for the mitogenome dataset and without partitions for the UCE dataset.We included previously published Neotominae and Sigmodontinae mitogenomes and UCEs downloaded from Gen-Bank; KY707304 and KY707303 (Sullivan et al. 2017) 1).Holochilus sciureus was used as an outgroup for both datasets.The BI analyses ran for 50 million generations sampling every 1,000 generations with a burn-in of 25%.Output parameters were visualized using Tracer v1.7.1 (Rambaut et al. 2018) to check for convergence between runs.FigTree v1.4.4 (http://tree.bio.ed.ac.uk/software/figtree/) was used to visualize the phylogenetic trees.All of these analyses were performed on the Smithsonian Institution High Performance Computing Cluster (Smithsonian Institution: https://doi.org/10.25572/SIHPC).
The UCE processing recovered 209,371 and 556,910 contigs for H. lophurus and Holochilus sciureus after Trinity assembly, respectively.A total of 3,654 UCE loci were obtained for H. lophurus and 2,186 for H. sciureus.The incomplete matrix contained 4,415 UCE loci (N = 20; average = 3,145; min = 1,087; max = 3,859).The 90% matrix included 438 UCE loci with an average of 15 informative sites and 649 bp length per locus.Our phylogenetic analyses support the placement of H. lophurus, H. simulatus, and H. ixtlani within a clade with high nodal support (Fig. 3).However, the clade comprising the genus Habromys is nested within a clade that contains members of the genus Peromyscus (sensu stricto; Carleton 1980).Therefore, our data support the hypothesis that Peromyscus is paraphyletic (Bradley et al. 2007;Miller and Engstrom 2008;Platt et al. 2015;Sullivan et al. 2017;Castañeda-Rico et al. 2022b).In addition, we corroborated previous results by Castañeda-Rico et al. (2022b, 2023) where members of the genus Habromys are sisters to P. aztecus, a member of the Peromyscus aztecus species group, and this clade is sister to a clade comprising members of the Peromyscus truei species group (P.attwateri and P. pectoralis).The difference between those previous studies and ours is that they only included H. simulatus in the UCE-based phylogeny and H. ixtlani in the mitogenome-based phylogeny (one sample for each dataset).In contrast, in this study, we have expanded our sampling and have included two species (H.simulatus and H. lophurus) in the UCE-based phylogeny and three species (H.simulatus, H. ixtlani, and H. lophurus) species (H. ixtlani, H. lepturus, and H. lophurus) (Carleton et al. 2002).Based on our limited sampling of species (3 out of 7) within Habromys, we cannot resolve the relationships within the genus or test the phyletic affinity among morphologically similar species.However, we show that using mitogenomes and nuclear genome-wide data provides well-resolved phylogenies that can be used to resolve or support opposite phylogenetic hypotheses.For example, León-Paniagua et al. (2007), using only three mitochondrial genes, found that H. lophurus is sister to H. ixtlani + (H.lepturus + H. chinanteco), followed by a sister relationship to H. delicatulus + H. schmidlyi, where H. simulatus is the most divergent species within the genus; however, Rogers et al. (2007), using one mitochondrial gene, found that H. lophurus is sister to a clade including H. ixtlani + H. lepturus, and this clade is sister to H. delicatulus, where H. simulatus + H. delicatulus are the most divergent of the genus, but they did not include H. schmidlyi in their study.Neither of these studies, based on limited data from mitochondrial genes, found support for the two clades suggested by Carleton et al. (2002), and they did not recover the same topology.In addition, some of the clades in those genetic studies were not highly supported.We expect that the mitogenomes and UCE loci generated in this study will contribute to future phylogenomic studies that aim to further elucidate the phylogenetic relationships and the evolutionary history of this critically endangered genus.
All Habromys species are of conservation concern given the increasing annual deforestation rates for montane cloud forest (Rogers et al. 2007;Contreras-Medina et al. 2022) as well as other characteristics described above.The International Union for Conservation of Nature (IUCN) established that populations of all Habromys are decreasing but only H. simulatus, H. ixtlani, H. schmidlyi, H. chinanteco, and H. lepturus are considered Critically Endangered (Álvarez-Castañeda et al. 2018a(Álvarez-Castañeda et al. , 2018b(Álvarez-Castañeda et al. , 2018c(Álvarez-Castañeda et al. , 2018d;;Vázquez 2018), while H. delicatulus is considered Endangered (Vázquez 2017) and H. lophurus Near-Threatened (Emmons and Vázquez 2019).However, only H. simulatus is listed in the Mexican Official Norm NOM-059-SEMARNAT-2010 (Secretaría de Medio Ambiente y Recursos Naturales 2010) under the category of Special Protection (Pr).The NOM-059 is the official document generated by the Mexican government that includes threatened animal and plant species.
Efforts to determine the correct status of these species have been stifled because of the very limited number of studies that have focused on this genus.Furthermore, a recent study looking at distributional patterns and conservation of the genus Habromys (Contreras-Medina et al. 2022) proposed that all Habromys species except H. lophurus must be included in the NOM-059 because their geographical distribution covers less than 5% of the Mexican territory based on the known distribution and their species-distribution models.However, they did not consider the particular case of H. lophurus and H. simulatus, where it has been proposed that they consist of two geographically isolated species (León-Paniagua et al. 2007).This urgently needs to be resolved in future studies and, if it is confirmed, it would provide additional support for the protection of these two geographically restricted and imperiled species as it would change the status of H. lophurus to an Endangered or Critically Endangered species.We propose that all species in this genus should be listed in the Mexican Official Norm under the category of Threatened (A) or Critically Endangered (P) and as Critically Endangered (CR) by the IUCN.

Figure 1 .
Figure1.Voucher specimen of Habromys lophurus (USNM 569392) from which the mitogenome and ultraconserved elements were obtained in this study.This species may be identified by its crested tail and usually by the absence of a white underside of the tail.Currently, there are no pictures available of live individuals of this species.Photograph by Katie Sayers.

Figure 2 .
Figure 2. Mitogenome map of the crested-tailed deer mouse Habromys lophurus (USNM 569392, GenBank accession number OP936043) depicting the gene organization.The inner circle shows the CG content along the mitogenome.

Table 1 .
Specimens examined in this study using mitogenomes and UCEs with species name, accession number collection/ID study, GenBank accession number, and GenBank BioProject.Smithsonian Institution's National Museum of Natural History (USNM), Museum of Texas Tech University (TTU/TK), Museo de Zoología "Alfonso L. Herrera" Facultad de Ciencias UNAM (MZFC), University of Michigan Museum of Zoology (UMMZ), and Museum of Vertebrate Zoology, University of California, Berkeley (MVZ).