The non-obese diabetic mouse sequence, annotation and variation resource: an aid for investigating type 1 diabetes

Abstract

Model organisms are becoming increasingly important for the study of complex diseases such as type 1 diabetes (T1D). The non-obese diabetic (NOD) mouse is an experimental model for T1D having been bred to develop the disease spontaneously in a process that is similar to humans. Genetic analysis of the NOD mouse has identified around 50 disease loci, which have the nomenclature Idd for insulin-dependent diabetes, distributed across at least 11 different chromosomes. In total, 21 Idd regions across 6 chromosomes, that are major contributors to T1D susceptibility or resistance, were selected for finished sequencing and annotation at the Wellcome Trust Sanger Institute. Here we describe the generation of 40.4 mega base-pairs of finished sequence from 289 bacterial artificial chromosomes for the NOD mouse. Manual annotation has identified 738 genes in the diabetes sensitive NOD mouse and 765 genes in homologous regions of the diabetes resistant C57BL/6J reference mouse across 19 candidate Idd regions. This has allowed us to call variation consequences between homologous exonic sequences for all annotated regions in the two mouse strains. We demonstrate the importance of this resource further by illustrating the technical difficulties that regions of inter-strain structural variation between the NOD mouse and the C57BL/6J reference mouse can cause for current next generation sequencing and assembly techniques. Furthermore, we have established that the variation rate in the Idd regions is 2.3 times higher than the mean found for the whole genome assembly for the NOD/ShiLtJ genome, which we suggest reflects the fact that positive selection for functional variation in immune genes is beneficial in regard to host defence. In summary, we provide an important resource, which aids the analysis of potential causative genes involved in T1D susceptibility. Database URLs: http://www.sanger.ac.uk/resources/mouse/nod/; http://vega-previous.sanger.ac.uk/info/data/mouse_regions.html#Idd

Figure 1.

Entry point to the Idd regions in Vega. The regions are represented graphically and shown in the relative position they are found in the C57BL/6J genome. Each region links through to a regional summary. The MHC annotation will be available in the resource by mid-2013.

Figure 2.

The NOD and C57BL/6J mouse sequences can be aligned against each other. Homologous genes are connected with lines to help identify them. Blocks of homologous sequence are coloured green, and regions with different sequence or no sequence are coloured light blue. It is clear that there are different intronic sequences present in gene Bcat1 in CHORI-29 (lower panel) with respect to C57BL/6J, possibly resulting in changes to regulatory regions or other functional sequences.

Figure 3.

Comparison of SNP sets in the NOD Idd regions obtained by the BAC sequencing and the MGP.

Figure 4.

Percentage of variants affecting each genomic element (introns excluded from the chart).

Figure 5.

Gene Bhlhe41 (yellow box) in the Idd6.1+2 region from GRCm38 C57BL/6J reference does not have a homolog annotated in NOD owing to a sequence gap (orange box). It is therefore not possible to be confident whether this gene is present and expressed in NOD mouse.

Figure 6.

A higher resolution view of gene Bhlhe41 taken from the Vega genome browser. RNA-Seq data from NOD has been uploaded into the browser and aligned to the GRCm38 C57BL/6J reference. This shows that the gene is clearly expressed in the NOD mouse. On closer inspection, it would appear that there may be evidence of a 3′ overlapping non-coding RNA locus supported by three mouse mRNAs from AK032333.1, AK040945.1 and AK079251.1 as illustrated by the yellow box with the blue outline. The ability to upload RNA-seq data provides a way to investigate gene expression for sequences not yet represented in the NOD Idd regions and could also prove useful in observing differential intergenerational gene expression.

Figure 7.

Analysis pipeline for the NOD mouse project. C57BL/6J genomic sequence in Idd regions is annotated before an annotation transfer using exonerate, shown here by the orange arrow. Transcript objects are then manually inspected again in the NOD mouse and further manual annotation carried out where appropriate. Unlike the C57BL/6J annotation, the NOD mouse annotation is only available in Vega.