Ramos Gonzalez, Daniel
(2021)
Colour polymorphism in the terrestrial snail Cepaea nemoralis: from genetics and genomics to spectroscopy and deep learning.
PhD thesis, University of Nottingham.
Abstract
Colour variation in the animal kingdom has been important in science to determine the principles of biology, especially in genetics and evolution. In the past decades, much effort has been targeted at the evolutionary, ecological and genetic basis of colour variation. Although land snails have been relatively neglected, especially in latter years, a comprehension of genetics and the evolution is important to understand colour variation precisely because snails may be representative of many species. When studying colour polymorphism, one of the remaining challenges is to describe colour. Generally, colour is described manually, relying on the judgment of human perception, classifying them into a discrete types. The main issue, then, is that human perception is subjective and colour is continuous. Fortunately, technology has enabled new techniques to score colour, which may help to investigate colour polymorphism.
This thesis aims to contribute to the knowledge of the maintenance of colour polymorphism by firstly, understanding the genetics and genomics and secondly, developing new methods for the scoring of colour. To achieve this, the grove snail Cepaea nemoralis was selected as a model species. Cepaea nemoralis was chosen due to their highly polymorphic shell, its easy collection, is widely distributed in all variety of habitats and the colour and banding morphs showing Mendelian inheritance (Cain & Sheppard, 1950, Cain & Sheppard, 1952, Cain & Sheppard, 1954, Lamotte, 1959, Jones et al., 1977).
In the first part, I aimed for a better understanding of the inheritance of colour. Hence, new crosses of C. nemoralis were used, with flanking restriction site–associated DNA sequencing (RAD-seq) markers used to identify putative instances of recombination with the supergene that determines colour and banding. No evidence of the predicted recombinants was found. Instead, a better explanation could involve incomplete penetrance and epistasis (Gonzalez et al., 2019). The findings therefore challenge the previous assumption of the supergene architecture and provides a new resource for the future creation of a fine mapping of the supergene (Gonzalez et al., 2019).
In the second part, I aimed to understand the evolutionary history of C. nemoralis, by investigating the relationship of the genomic and supergene variation with the geographic distribution over Europe. High-throughput genome-wide genotyping was achieved via a double digest restriction-site associated DNA sequencing (ddRADseq) method. A broad phylogenomic relationship showed geographic structure. However, no relationship between the geographical distribution and colour variation was found. Furthermore, possible genomic regions under selection, which may be driving the genomic variation, were identified. In addition, the phylogeny described the evolution of C. nemoralis and indicated how the Pyrenean lineages colonised Europe after the Pleistocene. The results suggest new roads of research into the evolutionary and genomic mechanisms that have led the geographical genomic and supergene variation of C. nemoralis.
In the third part, colour manual scoring was tested using new quantitative methods to describe colour to better understand colour variation. Therefore, a comparative study with historical and present shell colour patterns of C. nemoralis in the Pyrenees was used. Prior studies manually scored shell ground colour into three discrete colours; yellow, pink or brown. However, colour is continuous and the description of discrete colours may incur potential error and biased results. Thus, a quantitative method to score shell colour and to test manual scoring, comparing patterns of C. nemoralis shell colour polymorphism was used. Similar altitudinal trends irrespective of the method were found, even though quantitative measures of shell colour reduced the possibility of error. Moreover, a remarkable stability in the local shell patterns over five decades were found. This study determined that both methods remains valuable illustrating several advantages and disadvantages. In the future, a combination of both methods may be a possible solution.
Finally, and as continuation of the third part, a new visual recognition and classification method for C. nemoralis based on spectrophotometry and deep learning was created. Firstly, colour of the shells were quantified by spectrometry, and secondly, pictures were taken of the measured shells, in different backgrounds. Those pictures were used to train and test a Region-based Fully Convolutional Networks (R-FCN). Furthermore, public domain pictures were collected from iNaturalist database (https://www.inaturalist.org/), to validate the model. The results illustrate that this method can achieve high accuracy of detection and classification of snails into the right morph. This work may facilitate the way of how colour polymorphism was investigated, illustrating new avenues for future research.
In conclusion, this thesis evaluates the limitations found in prior studies and generates new data for the genetic and genomic understanding of C. nemoralis colour polymorphism. It also produced viable solutions, using new technologies, to score the diverse colour morphs. I also contributed to the geographic evolutionary genomic diversity knowledge.
Actions (Archive Staff Only)
 |
Edit View |
Tools
Tools
