Despite over 100 years of study, the location of the fully sex-linked region of the guppy (Poecilia reticulata) carrying the male-determining locus, and the regions where the XY pair recombine, remain unclear. Previous population genomics studies to determine these regions used small samples from recently bottlenecked captive populations, which increase the false positive rate of associations between individuals' sexes and SNPs. Using new data from multiple natural populations, we show that a recently proposed candidate for this species' male-determining gene is probably not completely sex-linked, leaving the maleness factor still unidentified. Variants in the chromosome 12 region carrying the candidate gene sometimes show linkage disequilibrium with the sex-determining factor, but no consistently male-specific variant has yet been found. Our genetic mapping with molecular markers spread across chromosome 12 confirms that this is the guppy XY pair. We describe two families with recombinants between the X and Y chromosomes, which confirm that the male-determining locus is in the region identified by all previous studies, near the terminal pseudo-autosomal region (PAR), which crosses over at a very high rate in males. We correct the PAR marker order, and assign two unplaced scaffolds to the PAR. We also detect a duplication, with one copy in the male-determining region, explaining signals of sex linkage in a more proximal region.

The sex chromosome pairs of many species do not undergo genetic recombination, unlike the autosomes. It has been proposed that the suppressed recombination results from natural selection favouring close linkage between sex-determining genes and mutations on this chromosome with advantages in one sex, but disadvantages in the other (these are called sexually antagonistic mutations). No example of such selection leading to suppressed recombination has been described, but populations of the guppy display sexually antagonistic mutations (affecting male coloration), and would be expected to evolve suppressed recombination. In extant close relatives of the guppy, the Y chromosomes have suppressed recombination, and have lost all the genes present on the X (this is called genetic degeneration). However, the guppy Y occasionally recombines with its X, despite carrying sexually antagonistic mutations. We describe evidence that a new Y evolved recently in the guppy, from an X chromosome like that in these relatives, replacing the old, degenerated Y, and explaining why the guppy pair still recombine. The male coloration factors probably arose after the new Y evolved, and have already evolved expression that is confined to males, a different way to avoid the conflict between the sexes.

16Introduction: 908 Abstract 33We describe new genetic mapping results from 6 full-sib families in the guppy (Poecilia 34 reticulata), two of which included recombinants between the X and Y chromosomes. These 35 recombinants confirm that the guppy sex-determining locus is in the region identified by all 36 previous studies, including a recent report suggesting a candidate sex-determining gene in 37 this fish, close to the pseudo-autosomal region (or PAR) at the chromosome terminus. Our 38 results suggest the presence of some errors in the current assembly of the guppy genome. 39In males, crossing over occurs at a very high rate in the PAR, and our genetic map of the 40 region allows us to correct the marker order. We also identified two unplaced scaffolds 41 carrying genes that map to the PAR. Genetic mapping cannot be used to order markers in 42 the region where crossing over is infrequent. However, our recombinant male is informative 43 about the order, under the reasonable assumption that crossovers are infrequent. Our 44 mapping families and natural population samples also show that the recently proposed 45 candidate for this species' sex-determining gene is not completely sex-linked. We detect an 46 association between individuals' sex and an SNP in the sex-determining region, but not with 47 a marker 0.9 Mb away from it, suggesting that variants in this region may be in linkage 48 disequilibrium with the actual sex-determining factor, but that the factor itself has not yet 49 been identified. So far, no consistently male-specific variant has been identified in the guppy 50 sex-determining region. 51 52 53

Genetic and physical mapping of the guppy (P. reticulata) have shown that recombination patterns differ greatly between males and females. Crossover events occur evenly across the chromosomes in females, but in male meiosis they are restricted to the tip furthest from the centromere of each chromosome, creating very high recombination rates per megabase, as in pseudo-autosomal regions (PARs) of mammalian sex chromosomes. We used GC content to indirectly infer recombination patterns on guppy chromosomes, based on evidence that recombination is associated with GC-biased gene conversion, so that genome regions with high recombination rates should be detectable by high GC content. We used intron sequences and 3rd positions of codons to make comparisons between sequences that are matched, as far as possible, and are all probably under weak selection. Almost all guppy chromosomes, including the sex chromosome (LG12), have very high GC values near their assembly ends, suggesting high recombination rates due to strong crossover localisation in male meiosis. Our test does not suggest that the guppy XY pair has stronger crossover localisation than the autosomes, or than the homologous chromosome in the close relative, the platyfish (Xiphophorus maculatus). We therefore conclude that the guppy XY pair has not recently undergone an evolutionary change to a different recombination pattern, or reduced its crossover rate, but that the guppy evolved Y-linkage due to acquiring a male-determining factor that also conferred the male crossover pattern. We also identify the centromere ends of guppy chromosomes, which were not determined in the genome assembly.