Abstract

The work presented here is one facet among a larger set of analyses underway with various collaborators that comprise the International Cat Genome Sequencing Consortium. The analyses came about after completing a recent assembly of the cat genome. I highlight our variant calling pipeline, as implemented at the McDonnell Genome Institute. The process of cats evolving from ancestral wildcats around 10,000 years ago stamped genetic signatures in their DNA. By comparing the domestic cat genome with whole genome sequences from wildcats, we began the work of pinpointing the genetic changes that drove the transformation from a wild population to a domesticated population, and this genetic evidence has allowed us to better understand feline biology within the broader context of domestication.

Introduction

The State of Cat Genomics

The National Human Genome Research Institute (NHGRI) supported a domestic cat (Felis catus) genome sequencing project as part of a larger effort to sequence mammalian genomes. To justify this effort, an improved feline genome assembly would assist in identifying disease-causing mutations for simple and particularly complex disease processes within cats. At the same time, with over 250 hereditary disorders, some of which are similar to genetic pathologies in humans, the domestic cat would serve as a model for human disease. Some examples include feline leukemia virus (FeLV) and feline immunodeficiency virus (FIV).

In the 2005 Proceedings for this conference, Dr. Kerstin Lindblad-Toh reported on the progress surrounding the cat genome:
"The genome sequence of a female cat will be ready in early fall of 2005. For this project each position has been sampled only ~ 2 times. Thus we expect the sequence to cover only ~ 80% of the cat genome. The genome will be compared to the other mammals to identify genes and other functional elements. A small effort to study the variation within the cat population is being discussed."

Within a year, the International Cat Genome Sequencing Consortium submitted that draft assembly to NCBI. In the decade since, we have provided two significant updates to the cat genome, using genetic material derived from the original female cat. The first update, submitted in 2011 and referred to as Felis_catus-6.2, is a chromosomal assembly comprised of 2.35 billion nucleotide bases. These positions were sequenced to an average depth of 14-fold (i.e., 14x). We used transcriptome data to build gene models for this version of the assembly and identified almost 19,500 protein-coding genes and nearly 1,900 non-coding RNAs. The analyses presented below relied on Felis_catus-6.2. The second update, submitted last year and referred to as Felis_catus-8.0, added 200 million bases of sequence and eliminated gaps that existed in v6.2, thereby improving the contiguity of the full assembly.

History of Cat Domestication

Archaeologists have found evidence of cats coexisting among early farming settlements.^1,2 Some hypothesize that farmers rewarded cats that protected their granaries from rodents. In a sense, early agricultural communities provided food in order to entice cats to remain even after their main prey, rodents, became depleted. Farmers then began selecting docile cats with desirable physical characteristics, which most believe led to the start of domestication. Physical and behavioral changes likely occurred simultaneously and began to shape the feline genome toward domestication.

For historical context, here is a simplified timeline of cat evolution:

 6MYA: The species members of Felis emerged. DNA changes resulted from millions of years of natural selection related to feline survival. Cats developed a unique need for a strict diet of meat along with enhanced senses for social communication and hunting for their main prey, small rodents.

 10KYA: Cats evolved from wildcats. The initiation of cat domestication probably began earlier as humans and ancestral cats became more and more interdependent, especially for controlling pests such as rodents and their associated diseases. Farmers rewarded cats that protected their grain stores from rodents with food. As farmers began selecting cats for docility and desirable characteristics, it effected physical and behavioral changes that began to shape the feline genome.

 150 Years Ago: Modern cat breeds evolved. Certain breeds of cats were artificially selected in various regions of the world. This process was likely based on aesthetics for preferred coat color and patterning, unlike dog breeds, which were selected for traits such as herding, hunting and protection.

The genetic basis of artificial selection underlying domestication is of great interest and still mainly unexplored. Little is known about the genetic changes associated with the process of cat domestication. Still less is known about the process of cat domestication relative to other carnivore genomes, including the domestic dog. Some would argue that housecats are only semi-domesticated, since they retain the proficient hunting behaviors that we see in wildcats. While this is open to debate, we nonetheless predicted a relatively modest effect of domestication on the cat genome based on recent divergence from wildcats and the lack of clear morphological and behavioral differences from wildcats, with docility, gracility, and pigmentation being the exceptions.

Methods

To identify genomic regions showing signatures of selection influenced by the domestication process, we used whole-genome analyses of cats from different domestic breeds and wildcats using pooling methods that control for genetic drift. Other genomic studies of domesticated animals used similar strategies, pooling samples from domesticated populations and comparing the genetic variation with their wild progenitors. The pooling approach is hypothesized to distinguish selective sweeps from genetic drift. As a result of this study,³ we made publically available the genomic sequences from various breeds as well as a pool of wildcat sequence data.

We sampled 22 cats from six diverse breeds that span the putative phylogenetic tree of domestic cats. The selected breeds included Birman, Turkish Van, Japanese Bobtail, Egyptian Mau, Norwegian Forest, and Maine Coon. The DNA samples were pooled by breed and sequenced to a range of depths, depending on the number of individuals per breed. The sequence data were then pooled for further analysis, with an aligned coverage depth for this domestic cat pool totaling approximately 55x. Two wildcats (Felis silvestris lybica) were sampled from Israel and Kazakhstan and another pair of European wildcats were trapped and sampled from France and Portugal. These samples were pooled and sequenced to an estimated depth of 7x.

Our bioinformatics pipeline separately aligned the pair of pooled sequences (i.e., the domestic pool and the wildcat pool) to the 6.2 reference assembly using a standard aligner, known as bwa. We called single nucleotide variants (SNVs) using two detection methods (VarScan and SAMtools) and used only those variants that were called by both methods. In order to account for potential inaccuracies in the reference assembly, we applied a clustered variants filter that allowed no more than five variants per 500-bp window. Once we settled on a final, high-confidence set of variants in both pools of data, we performed standard calculations that are typically used in population genetics: 1) F_ST and 2) heterozygosity, both of which were calculated along 100 kilobase (kb) sliding windows with a 50-kb step size. We calculated F_ST as a metric of genetic divergence between the domestic cat pool and the wildcat pool, and we calculated heterozygosity (Hp) as a metric of genetic diversity within the domestic pool and within the wildcat pool. Hp was calculated for both pools.

The underlying theory to this approach was to identify loci that showed a relatively low proportion of variation within the domesticated population while also showing a relatively high proportion of variation between the domesticated and wild population. We Z-transformed the results and selected as outliers only the genomic windows that fell beyond four standard deviations from the mean. Finally, we annotated for gene content within these outlier genomic windows. In summary, we determined which of the domestic cat genes showed low diversity in domestic cats and high divergence from wildcats and postulated that these particular genes were under the heaviest selection during the transition from wild to domestic. We also only focused our attention on the windows with coding sequence. In the future, a more thorough analysis on the regulatory regions might uncover additional genomic regions that do not encode for proteins, but rather regulate their expression.

Results

Candidate Domestication Regions in the Domestic Cat

One of the first questions we asked was whether any of our candidate genes overlapped with those discovered in a recent analysis that compared a pool of domestic dog breeds with a population of wolves.⁴ The answer is no. The dog analysis described 35 regions in the dog genome with high F_ST and low Hp that overlapped with 122 genes. As described below, we uncovered five regions that spanned twelve known genes and one unknown gene. There were no shared genes within the dog and cat candidate domestication regions.

We identified twelve genes that changed as cats became domesticated. Some of these, based on previous studies of knockout mice, seem to play a role in cognition and behavior, including fear responses and the ability to learn new behaviors when given food rewards. We see two regions located on chrA1, two regions on chrB3, and a large region on chrD3. A subset of genes may elicit further examination with more detailed and functional studies. For instance, the protocadherin genes from chrA1 have implications for neuronal connections and fear conditioning in mice. The glutamate receptor is one of the more common receptors in the mammalian brain, playing a role in memory formation and reward learning. DCC, along chrD3, also has implications for reward responses in mice. Given the gene product's interaction with netrin, which plays a role in neural guidance for both axon outgrowth and repulsion, we propose that this gene is good candidate for continued analysis. A fascinating prospect about protocadherins, glutamate receptors, and DCC involves their implications in other mammalian domestication stories, including dogs,⁵ horses,⁶ pigs,⁷ and rabbits⁸. An important question for future research will be if specific genetic variants, or genetic variants within specific types of gene families, contribute to the distinctive behavioral or physiological traits across different domestication events.

We also found putatively selected genes in domestic cats that influence the migration of neural crest cells, which are stem cells in developing embryos affecting everything from skull shape to coat color. This finding supports a recent proposal⁹ that such cells may act as a sort of master control switch of domestication and potentially explains why domestic animals share common traits, such as smaller brains and certain pigmentation patterns - a mystery first noted by Charles Darwin.

Domestic Cats Compared with Other Mammals

We next compared the domestic cat genome to other mammals, including human, tiger, dog, and cow. We started with the genome of a domestic cat and identified genes that showed a high degree of similarity with those of tigers, dogs, humans, and cows. Our analysis of the Felinae lineage revealed 281 genes that showed signs of rapid or numerous genetic changes - a hallmark of recent (i.e., positive) selection - in domestic cats. Amino acid changes in many of the cat genes were shown to contribute to significant structural or biochemical effects. Some of these genes were involved with hearing and vision, the senses that felines heavily rely on. We found at least six genes under positive selection in cats that were associated with hearing capacity; we know this because mutations in these genes cause non-syndromic recessive hearing loss or deafness. At least twenty genes under positive selection in carnivores were associated with vision-related pathways, which fit with the importance of visual acuity for these natural-born hunters. Other genes play a role in fat metabolism and are likely an adaptation to cats' highly carnivorous lifestyle. Cats rely less on their sense of smell for hunting than dogs do, which was apparent from the smaller repertoire of functional olfactory receptor genes in the feline genome. However, the cat genome harbors a higher proportion of functional genes encoding vomeronasal sensation. The vomeronasal organ is a sort of auxiliary sense of smell, mainly used to detect pheromones. It's been suggested that there's a tradeoff between olfactory and vomeronasal capacity in evolution, and the cat's genome supports that.

Genetics of Coat Coloration in the Birman Breed

White coloring in general is trait that has been linked to domestication. A recent study in several white-spotted cat breeds localized the mutation responsible for the spotting pigmentation phenotype within the first intron of the gene, KIT.¹⁰ This gene, located on cat chromosome B1, is primarily involved in melanocyte migration and survival. Surprisingly, direct genetic sequencing excluded the published dominant allele as being associated with the white coloration pattern in the Birman cat breed. Targeted sequencing on a larger group of domestic cats (409 from 21 breeds, 5 Birman outcrosses, and 315 random bred cats) made it possible to narrow in on novel, recessive mutations in the KIT gene that appear to explain the characteristic white feet found in Birman cats. We identified just two adjacent missense mutations that were concordant with the gloving pattern, and these mutations possibly contribute to producing the white gloving pattern in the paws of this particular cat breed.

Conclusions

We have begun to pinpoint the genetic changes that drove the transformation to the domestic cat. Our initial methods reveal 5 regions that harbor 12 known genes, some of which are implicated in neural processes and neural-crest cell migration. Collectively, in searching for insight into the process of cat domestication, these findings provide working hypotheses for continued sequencing efforts (of both domestic breeds and wildcats). Finally, continued improvement of the domestic cat reference genome, along with improved gene predictions, will advance our knowledge of cat biology and domestication events in mammals.

References

1.  Hu Y, et al. Earliest evidence for commensal processes of cat domestication. Proceedings of the National Academy of Sciences USA. 2014;111(1):116–120.

2.  Vigne J-D, et al. Early taming of the cat in Cyprus. Science. 2004;304(5668):259.

3.  Montague MJ, et al. Comparative analysis of the domestic cat genome reveals genetic signatures underlying feline biology and domestication. Proceedings of the National Academy of Sciences USA. 2014;111(46):17230–17235.

4.  Axelsson E, et al. The genomic signature of dog domestication reveals adaptation to a starch-rich diet. Nature. 2013;495:360–364.

5.  Li Y, et al. Domestication of the dog from the wolf was promoted by enhanced excitatory synaptic plasticity: a hypothesis. Genome Biology and Evolution. 2014;6(11):3115–3121.

6.  Schubert M, et al. Prehistoric genomes reveal the genetic foundation and cost of horse domestication. Proceedings of the National Academy of Sciences USA. 2014;111(52):E5661–E5669.

7.  Moon S, et al. A genome-wide scan for signatures of directional selection in domesticated pigs. BMC Genomics. 2015;16(1):130.

8.  Carneiro M, et al. Candidate genes underlying heritable differences in reproductive seasonality between wild and domestic rabbits. Animal Genetics. 2015;46(4):418–425.

9.  Wilkins AS, Wrangham RW, Fitch WT. The "domestication syndrome" in mammals: a unified explanation based on neural crest cell behavior and genetics. Genetics. 2014;197(3):795–808.

10. David VA, et al. Endogenous retrovirus insertion in the KIT oncogene determines white and white spotting in domestic cats. G3: Genes Genomes Genetics. 2014;4(10):1881–1891.