Unwinding snail chirality by a massvie subtactive linkage analysis (MSLA)

The aim of this project is to utilise the power of ultrahigh-throughput DNA sequencing to directly isolate the gene product that determines chirality in the snail Lymnaea stagnalis, by a method that we term 'massive subtractive linkage analysis' (MSLA), a new application for a the technology. The basic principle is that if the dextral gene product is present in a pool of RNA from snails that are genetically dextral, and the sinistral gene product is present in the RNA of snails that are genetically sinistral, then a comparative bioinformatic analysis of the sequence reads can be used to rapidly identify candidate genes. With false positives excluded by genetic mapping (using SNPs), we will then attempt to definitively identify the gene with functional and cytological studies. The steps by which the main objective will be reached are: 1. A dextral snail (DD) will be crossed with a sinistral (SS) to create an F1 that is heterozygote (DS) for the chirality locus. This individual will then be allowed to reproduce by self-fertilisation, and a large F2 generation raised to adulthood. The chirality genotype of individual F2 snails will then be determined, by scoring the phenotype of their offspring. With an F1 of genotype DS, then the expectation is that F2s will be of genotype DD or DS (yielding dextral babies) or SS (sinistral babies) in a ratio of 3:1. 2. 20 to 40 F2 snails of different chirality genotypes will be selected, and tissue removed from their ovotestes (an ovotestis is the combined ovary and testis of a hermaphrodite). Two separate pools will then be made by combining together tissue samples from snails of chirality genotype DD/DS, distinct from another pool made from snails of genotype SS. Messenger RNA will then be extracted and cDNA synthesised. cDNA from the two pools (DD/DS versus SS) will then be sequenced to completion using the 454 Life Sciences facility at Liverpool. 3. As a result of recombination during meiosis in the F1, individual F2 snails of different chirality genotypes will on average only differ at the chirality locus, or else linked genes. A bioinformatic comparison of the assembled sequence reads from the two pools (genetically dextral snails versus sinistral snails) will therefore allow rapid identification of the candidate dextral allele. 4. Single nucleotide polymorphisms (SNPs) will be identified for candidate genes. A larger sibship will be used to follow the segregation of these markers in the F2: if a candidate is a transcript of the chirality locus, then SNPs will segregate perfectly in the F2. If the SNP genotyping is unproductive, then this could be because copy number alone determines the phenotype. In this circumstance, we will use quantitative PCR to assess ovarian transcript levels of RNA in a segregating sibship of F2 snails. The expectation is that the chirality allele should be at the same relative level in dextral versus sinistral snails. 5. To finally confirm the identity of the chirality gene product, we will attempt an assay for gene function, concentrating first on the most likely gene candidates. The full length cDNA clone will be retrieved by RACE-PCR, cloned into an expression plasmid, mRNA/protein synthesised, then injected into a 1-cell zygote. The expectation is that the chirality of a sinistral snail will be transformed by the dextral allele, but the reciprocal experiment will have no effect. To support the functional assay, cytological studies of gene expression and other association studies will also be carried out.

Much of the biological world has "handedness" - a differentiation of left and right. This pattern is found in the basic molecules of life, many of which come in pairs of mirror-image varieties, all the way through to our own bodies. Animals are usually apparently symmetrical across the long body axis (head to tail), but this superficial symmetry hides fundamental differences between left and right (such as the placement of out hearts). In snails, the breaking of symmetry is even more clear, as their spiral shells must curl to the right (dextral) or left (sinistral). In the pond snail Lymnaea stagnalis animals can be left- or right-handed coilers, and this trait is inherited through their mothers. Oddly for a genetic trait, it is the mothers genes that determine whether her offspring are left- or right-coilers. We have been working to identify the gene or genes that work in the mother snail to determine her childrens' coiling. Using a new application of high-throughput DNA sequencing we have narrowed down the location of the handedness gene to only 400,000 of the 1 billion DNA letters of the snail genome. Very soon we will have a complete map of these 200,000 letters and we will be able to identify which of the genes in this region underpins the coling trait. From this we will be able to work out how left-handed snails become different from their right-handed siblings. As all animals have handedness, we also hope that this mechanism identified in snails will also be operating in many other kinds of animals.

(a) genome sequence survey of Lymnaea stagnalis (30x coverage in 100 base paired end Illumina reads), preliminary assembly, annotation)
(b) transcriptome survey of Lymnaea stagnalis embryos (200000 454 reads, assembly, annotation)
(c) RAD-Seq data for Lymnaea stagnalis cross (20 individuals, 20 M reads and pairs, RAD locus predictions, paired end assemblies)
(d) BAC sequences from chirality locus of Lymnaea stagnalis (complete sequences of BAC clones, assemblies and annotations)
(e) RADtools software suite for analysis of RAD data including paired end assembly
(f) RAD adapters: the RAD adapters developed for this work have been distributed widely in the UK research community
This grant aimed to identify the genetic locus underpinning chirality in the pond snail Lymnaea stagnalis by using an unique set of dextral and sinistral lines of snails maintained and inbred by Davison in Nottingham, the power of next-generation sequencing and bioinformatic analysis built by Blaxter in Edinburgh to identify polymorphic markers in these populations, the mapping of the populations and the chirality locus in a series of controlled crosses by Davison and Aboobaker in Nottingham, and functional exploration of candidate loci found by Aboobaker in Nottingham. This was to be effected by contributions from all three PIs and two PDRAs - a full time molecular geneticist in Nottingham and a 60% bioinformatician in Edinburgh.
We were fortunate in Edinburgh to recruit Dr John Davey, a computer scientist with a PhD in neuroscience/neuroinformatics. John is a very able and engaging researcher, and has contributed hugely to this project. The remainder of PDRA Davey's salary has been sourced from additional research grant funding on closely related projects from NERC and BBSRC.
The 18 months that elapsed between the conception and submission of the grant application and the start of the work spanned a revolutionary change in genomics as applied to neglected species. While we originally planned to use cutting edge technology to identify a large number of potential markers in exonic regions of the snail genome, and then to build platforms to score these polymorphisms in families segregating for chirality, the publication in late 2008 of the first next-generation Restriction-site Associated DNA sequencing paper from the Cresko and Johnston labs in Oregon changed the field. We were the first project in the UK to move to use RAD-Seq for population and developmental genetic analyses, and thus PDRA Davey’s role changed from the day he arrived to development of tools to analyse and interrogate RAD-Seq data (while the Nottingham PDRA developed RAD-Seq wet lab protocols). Edinburgh produced the adapters needed for generation of RAD-Seq libraries, and supplied these to Nottingham (indeed these adapter sets have now been used across the UK). Edinburgh also performed extensive troubleshooting of the RAD-Seq process, including optimisation of loading and data generation on the Illumina instruments and primary data QC.
PDRA Davey developed a suite of tools for RAD-Seq analysis, RADtools, and these have been released openly through a dedicated web site (https://www.wiki.ed.ac.uk/display/RADSequencing/Home). They remain the toolset of choice for many RAD-Seq applications. PDRA Davey also collaborated closely with the Oregon team in analyses, in particular with Julien Catchen, the author of the STACKS RAD-Seq analysis pipeline.
As generation of the L. stagnalis libraries took some considerable time, PDRA Davey collaborated with other UK researchers in analysis of RAD-Seq data, notably Simon Baxter of Cambridge, and a team based across Scotland working on genomic linkage analyses in farmed salmon. These projects served to tune the analytical pipeline and test it on real-world complex data.
Data have been generated for three different L. stagnalis samples. The first is a ~30 fold coverage of the ~1 gigabase L. stagnalis genome in Illumina paired end 100 base reads. These data have been assembled to yield a first pass genome sequence estimate for L. stagnalis. While the assembly is necessarily very fragmented, it does allow robust mapping of RAD-Seq and RNA-Seq reads (see below) and thus provides a bedrock for analysis. These data have not yet been made public (the raw data are in submission to ERA), and will form part of the first collaborative publication from the project.
The second data set is Roche 454 RNA-Seq from early embryos of L. stagnalis. The specification of whether the soma of an embryo is left or right coiled is made by the mother, and first expressed in the very early embryo (visible as chirality in the twisting of the upper and lower four-cell sets in the eight cell embryo). We generated over 200,000 reads and have used these to build an estimate of the transcriptome of these embryos. In the data we can identify both core metabolic genes as well as those we would expect to be involved in developmental regulation (i.e. candidates for the expression of the chirality phenotype, likely downstream of the causative controlling locus).
The third set is RAD-Seq data from a pedigree of inbred snails painstakingly constructed by Davison in Nottingham. By choosing twenty individuals, ten with the dextral phenotype (and genetically either DD or Dd; dextrality is dominant) and ten with the sinistral phenotype (dd), offspring of the same self-fertilised Dd snail. This snail was from the inbred line selected by Davison by selecting sinistral-carrying offspring of backcrosses to dextral parents. We generated 18.5 million paired-end 100 base RAD tag sequences, and using RADTools transformed these into 62,758 candidate alleles at 52,154 candidate loci. We identified 20 loci with polymorphisms that were in complete linkage with the chirality locus (these were absent in all sinistral snails but present in genetically dextral snails). Using paired-end assembly we identified contigs for each of these loci and developed PCR assays for each allele. These have been tested across over 2200 additional snails by the Nottingham PDRA, and we now have a good linkage map of the region of the genome spanning the chirality locus. Excitingly two of the markers clearly bracket the locus, with few recombinations mapped to the interval between them in 2000 snails (RAD locus "E" is