Extant and extinct bilby genomes combined with indigenous knowledge improve conservation of a unique Australian marsupial

The Ninu (Greater bilby, Macrotis lagotis) is a desert-dwelling, culturally and ecologically important marsupial. In collaboration with Indigenous rangers and conservation managers, we generated the first Ninu chromosome-level genome assembly (3.66 Gbp) and genome sequences for the extinct Yallara (Lesser bilby, Macrotis leucura). We developed and tested a scat SNP panel, based on our genomic datasets, to inform current and future conservation actions, to undertake future ecological assessments, and improve our understanding of Ninu genetic diversity in managed and wild populations. We also assessed the beneficial impact of targeted conservation actions, like translocations, in the contemporary metapopulation (N=363 Ninu). Resequenced genomes (temperate Ninu=6; semi-arid Ninu=6; Yallara=4) revealed two major population crashes during global cooling events for both species and differences in Ninu genes involved in anatomical and metabolic pathway adaptations to aridity. Despite their 45-year long captive history, Ninu have fewer long runs of homozygosity than other larger mammals, which may be attributable to their boom-bust life-history. We also investigated the unique Ninu biology using 12 tissue transcriptomes revealing expression of all 115 conserved eutherian chorioallantoic placentation genes in the uterus; an XY1Y2 sex chromosome system generated by fusion of the X with a large telocentric autosome; and expansions in olfactory receptor genes. Together, we demonstrate the holistic value of genomics in improving key conservation management actions, understanding unique biological traits, and developing tools for Indigenous rangers to monitor remote wild populations.

Methods
There are two datasets included in the Excel workbook:

A set of SNPs (n=9906) generated using DArTseq, a form of reduced representation sequencing and used in the population genetic analyses of the Ninu (n=363 samples). See Supplementary Note 2.4 for details of SNP calling and filtering. Briefly, reads were cleaned and aligned to the Ninu reference genome generated in this study (v1.9). Variants were called with Stacks and filtered to retain SNPs with a minor allele frequency of ≥0.01, minimum average allelic depth of 2.5 x per allele, allelic coverage difference ≤80%, call rate ≥70%, locus heterozygosity ≤90%, and reproducibility between technical replicates ≥90%, and remove putatively sex-linked SNPs.
A set of SNPs (n=35) generated using DArTseq, a form of reduced representation sequencing and selected for use in the MassARRAY Ninu scat genotyping assay. The DArTseq reads were mapped to an earlier version of the Ninu reference genome generated in this study (v1.4.3). See Supplementary Note 2.5 for details of the MassARRAY SNP Panel Design.

The .txt file contains the R code in Rmarkdown format used to select SNPs for the scat genotyping assay.

The file also contains a README with additional information about the coding of genotypes, and a sample key that contains population metadata for each genotyped sample.