Highlighting the Importance of Correct Sex Identification in Chondrichthyan Genomic Studies, Using the White Shark as an Example

Abstract

The increased availability of reference genome assemblies of sharks and rays has contributed greatly to our understanding of their biology, including their sex-determination mechanisms. However, several publicly available genome assemblies of sharks and rays appear to be missing information about the sex of the source individuals. This can confound the investigation into genetic sex-determining elements and hinder the discovery of sex-specific patterns. Herein, we highlight the importance of clear and accurate sex identification in sharks and rays for future genome assemblies, using an example of a white shark (Carcharodon carcharias) genome, in which the phenotypically assigned sex conflicts with the genetic information. This genome assembly was reported to be sourced from a juvenile female (BioSample: SAMN01915239). We analyzed the assembly by mapping its available genome sequences to the current white shark reference genome assembly and compared the read coverage to sequences collected from other samples. Evidence suggests that this specimen is genetically male, which contradicts its assignment based on phenotype. Therefore, we urge researchers to provide as much accurate information (e.g., sex, sampling localities, and life history) as possible when publishing genome assemblies for sharks and rays (or for any other organism).

1. Introduction

The advent of high-throughput sequencing is transforming biology. It is now possible to assemble high-quality reference genomes across the tree of life for species of interest. These reference genomes allow us to better understand the processes involved in development [1,2], the genetic basis of diseases [3,4], the evolution and diversification of life on our planet [5,6,7,8], and the conservation status of endangered species [9,10,11]. However, establishing a reference genome requires considerable time, effort, and resources [12]. Therefore, individuals from which a DNA sample is collected for reference genomes should ideally be well characterized to the greatest extent possible, including the individual’s developmental stage and sex.

Accurate sex annotation for samples enables studies that inform sex-specific migratory or demographic history, which can then be applied to inform effective conservation status and practices in the species of interest [13,14,15]. Genome assemblies have been demonstrated to bring insights into the sex determination of sharks and rays (e.g., [16,17]). However, a significant proportion of the existing shark and ray genome assemblies were sourced from samples whose sexes were undetermined or were not provided (Figure 1, Table S1), as sex is not always the focus of genomic studies. Indeed, studies of a reference genome assembly with unknown sex have brought great understanding into elasmobranch molecular evolution (e.g., the cloudy catshark Scyliorhinus torazame [2]). However, in species with heterogametic sex chromosomes, a lack of prior knowledge of the sex of the individual may cause misassembly in the sex-linked regions and may potentially propagate the errors into downstream analyses. For example, if the X and Y chromosomes are collapsed into one scaffold in a genome assembly, mapping short-read sequences from an individual of heterogametic sex (i.e., male) may display elevated nucleotide diversity and heterozygosity in these regions. The potential consequences of such a mistake may be indirectly observed in studies where elevated nucleotide diversity was used to infer non-recombining genomic regions [18,19]. Without the expectation of heterogametic chromosomes, as in these studies, these observations alone may lead to incorrect downstream interpretations of balancing selection and recombination rate.

Elasmobranch sex determination has been informed by cytogenetic studies and is generally observed as exhibiting a heteromorphic XX/XY genetic sex determination system (e.g., [20,21,22]). However, the reporting of X and Y chromosomes, to date, has been taxonomically limited [23], and the sex-determining genes have not yet been identified. In the process of exploring sex determination and the evolution of sex chromosomes in elasmobranchs, we used a publicly available white shark reference genome (Carcharodon carcharias) ([24], GenBank assembly GCA_003604245.1), mostly belonging to a female (BioSample SAMN01915239), and a more recent genome assembly of a male white shark, made available by the Vertebrate Genomes Project (VGP, https://vertebrategenomesproject.org/, accessed on 5 February 2021). The latter reference genome (GenBank assembly GCF_017639515.1, hereafter referred to as “VGP’s assembly”) was assembled through meticulous curation, and the heterogametic sex chromosomes (i.e., X and Y in white sharks) were identified, based on the half coverage and the homology of sequences to the sex chromosomes of other species [25].

Based on the examination of the two available white shark reference genome assemblies and the additional male and female genomic sequences that we collected, we report that the genomic DNA sequences of BioSample SAMN01915239, provided by Marra et al. [24] (hereafter referred to as “Marra et al.’s assembly female”), show a strong male signature, contrary to its original designation. This brings into question the likelihood of sex misidentification based on visual cues and potentially reflects the possible dynamism of sex determination in white sharks. Additionally, this uncertainty of sex identification highlights the importance of the thorough documentation of sexual traits, particularly in samples where genomic data are available, as future investigations can be facilitated by the vigilant monitoring of possible discordance between phenotypic and genotypic sexes.

2. Materials and Methods

We retrieved the Illumina sequencing reads for Marra et al.’s assembly female from SRA (https://www.ncbi.nlm.nih.gov/sra, accessed on 6 July 2023 and 11 September 2024; GenBank accession no. SRR1705785, SRR7693858, SRR7693861–SRR7693870) with the SRA toolkit v3.0.5 and v3.0.8 (https://github.com/ncbi/sra-tools, accessed on 9 May 2023 and 14 December 2023).

To compare the sequencing signatures of males and females, we further collected whole genomic short-read sequencing data for 5 male and 5 female white sharks sourced from four major ocean basins where the species is found (i.e., the East Pacific Ocean, Northwest Atlantic Ocean, West Pacific Ocean, and West Indian Ocean; see Table S2) to minimize the potential biases due to genetic variation that may have been introduced by population structure [26]. For each specimen, fin clips or muscle tissues were collected and stored in a 95% ethanol solution at 4 °C until the DNA extraction was performed.

DNA extraction from the tissue samples was conducted using an E.Z.N.A. Tissue DNA Kit (Omega Bio-Tek, Inc., Norcross, GA, USA), following the manufacturer’s protocols. Samples were delivered to the UF ICBR (Gainesville, FL, USA) for library preparation using the NEBNext Ultra II DNA Library Prep Kit for Illumina (New England Biolabs, Inc., Ipswich, MA, USA), and then sequenced through the Illumina Novaseq 6000 sequencing platform aimed at a coverage depth of ~20× for each individual (except for the 2 samples sequenced at ~70× coverage, see Table S2) based on the estimated genome sizes for each species, as noted on the genome pages of the NCBI datasets.

Adapter sequences and low-quality regions (<Q20) in all Illumina data were trimmed with Trim Galore v0.6.10 (https://github.com/FelixKrueger/TrimGalore, accessed on 5 July 2023). The trimmed reads were mapped to VGP’s assembly using BWA-MEM [6]; duplicate reads were marked and removed with Picard MarkDuplicates (accessed through the gatk package v4.4.0.0 [27]), followed by implementing a mapping quality filter through SAMtools v1.15-v1.20 [28], set at the minimum MAPQ = 60, to retain only the reads that were mapped to the reference at the minimum error rate for BWA-MEM. The paired-end reads acquired from the 10 samples were then pooled into the aggregated male and the aggregated female sequences, according to the assigned sexes. The sequencing depth on the Y chromosome of each dataset was calculated with SAMtools v1.20 [28] and then visualized with the ggplot2 package [29] in R v4.0.1 [30].

We also assessed the ratio of sequencing depth between the autosomes and the X chromosome to verify the sex assignments. In male heterogametic sex-determining systems, the average read depth ratio of an autosome to an X chromosome should be approximately 1:1 in females and 2:1 in males. We selected chromosome 32 (30,807,379 bp) and chromosome 34 (30,049,857 bp) as the representatives for the autosomes, as these two chromosomes are the closest in size to the X chromosome (28,814,942 bp) according to VGP’s assembly. The χ² test of the goodness-of-fit on the average read depth ratios was performed in R v4.0.1 [30].

3. Results

After adapter trimming and initial quality trimming with Trim Galore, 5,853,736,471 reads from the short-read sequences from Marra et al.’s assembly female, 4,440,606,223 reads from the aggregated male sequences, and 4,648,525,667 reads from the aggregated female sequences were mapped to the reference genome. After applying the stringent mapping quality filter (MAPQ = 60), the sequencing depth across the length of the Y chromosome for Marra et al.’s assembly female, aggregated male sequences, and aggregated female sequences are shown in Figure 2. The difference in sequencing depth for the Y chromosome is most conspicuous in the >2 Mbp section between the aggregated male (Figure 2a) and the aggregated female (Figure 2c) sequences. The sequencing depth of Marra et al.’s assembly female, when mapped onto the Y chromosome of VGP’s assembly (Figure 2b), has a nearly identical pattern to that seen for the aggregated male sequences (Figure 2a), including the patterns of depth variation due to the complex repeated sequences, a feature commonly observed in the chromosomes of other species where the X and Y chromosome sequences are highly diverged [25]. The sequencing depth of the aggregated female sequences (Figure 2c) at >2 Mbp is much lower, with an average depth of 0.213, contrary to 21.989 of the aggregated male sequences and 35.566 (27.612, after scaling the mapped reads to the average between the aggregated male and female sequences) of Marra et al.’s assembly female.

The first ~2 Mbp of the Y chromosome has a male-to-female sequencing depth ratio close to 1 and contains plenty of heterozygous sites in male samples, which is inconsistent with the supposed hemizygosity of a Y chromosome. Hence, this region resembles a putative pseudoautosomal region (PAR), a feature of the sex chromosomes that has previously been reported in the X chromosome of the white-spotted bamboo shark (Chiloscyllium plagiosum [16]) and zebra shark (Stegostoma trigrinum [17]).

When mapped to chromosome 32, chromosome 34, and chromosome X of VGP’s assembly, the average sequencing depths of Marra et al.’s assembly female on the three chromosomes show a closer pattern to the depth of aggregated males, rather than to the aggregated females (

Table 1. The average sequencing depths of reads from Marra et al.’s assembly female, the aggregated male sequences (n = 5), and the aggregated female sequences (n = 5) across the length of Chr 1, Chr 32, Chr 34, Chr X, and Chr Y, based on the current white shark reference genome assembly (GCF_017639515.1).

	Average Sequencing Depth
	Chr 1	Chr 32	Chr 34	Chr X	Chr Y
Marra et al.	163.597	138.003	118.415	58.770	34.778
Marra et al. (*)	127.009	107.139	91.932	45.626	27.000
Aggregated males	107.164	93.507	83.988	41.034	23.439
Aggregated females	111.161	96.347	86.014	80.575	10.273

*: Scaled coverage depth to the aggregated sexes, calculated by scaling the mapped overall read counts from Marra et al.’s assembly female to the average between the mapped overall read counts of the aggregated males and aggregated females.

, Figure 3). In both the aggregated males and Marra et al.’s assembly female, the averages and modal values of the sequencing depths on chromosomes 32, 34, X, and Y were close to 2:2:1:1 (excluding positions with coverage close to 0, caused by the stringent mapping quality filter), while in the aggregated females, the ratio is close to 1:1:1:0. For Marra et al.’s assembly female, the χ² test conducted on the goodness-of-fit of the sequencing depth ratio of Chr32:ChrX and Chr34:ChrX to a 2:1 ratio resulted in good fits (p-values = 0.302 and 0.963, respectively), while a fitting of 1:1 for Chr32:ChrX and Chr34:ChrX are soundly rejected (p-values of 1.620 × 10⁻⁸ and 7.435 × 10⁻⁶, respectively). As this result aligns with the goodness-of-fit results of the aggregated male sequences and opposes the aggregate female sequences (

Table 2. The χ² goodness-of-fit test p-values of an autosome to the Chr X ratios of the average sequencing depth of Marra et al.’s assembly female, the aggregated male sequences (n = 5), and the aggregated female sequences (n = 5) when tested against a male (2:1) and female (1:1) distribution.

	χ2 Test p-Values
	Male Expected Ratio 2:1		Female Expected Ratio 1:1
	Chr 32: Chr X	Chr 34: Chr X	Chr 32: Chr X	Chr 34: Chr X
Marra et al.	0.302	0.963	* 1.620 × 10–8	* 7.435 × 10–6
Aggregated males	0.486	0.903	* 6.072 × 10–6	* 1.222 × 10–4
Aggregated females	* 5.711 × 10–4	* 3.853 × 10–5	0.236	0.674

*: The value is significant, with the critical value set to 0.05.

), we conclude that the sequencing depth patterns of the two autosomes and the X chromosome corroborate our observation in the read coverage of the Y chromosome and that Marra et al.’s assembly female shows a strong signature of a male white shark genome assembly.

4. Discussion

Our understanding of the genomic structure in elasmobranchs has greatly benefited from the assembly of reference genomes, and the accurate assignment of sex is critical for research aimed at understanding sex determination. While certain elements of sample information may be less relevant to the original investigators, it can be crucial for future studies with different perspectives. Marra et al. [24] generated the first assembled reference genome for the white shark and adopted sophisticated analyses that brought insights into the maintenance of genome stability in sharks. Although we do not challenge these results, our analyses, based on a more recent high-quality reference genome, raise potential questions about the sex of the juvenile white shark from which Marra et al.’s genome assembly was sourced.

Interestingly, in white sharks, conflicting sex identifications between visual and genetic methods have been documented elsewhere. In designing sex-specific markers for white sharks, Devloo-Delva et al. [31] reported that 6.7% of the 402 white sharks examined demonstrated discrepancies between their genetic and visual sex assignments, which they attributed to potential visual identification errors or the presence of hermaphroditic individuals. While this discrepancy may equally raise doubts about the reliability of genetic-based sex identification or the practices of visual sex identification, it may be difficult to assume a widespread visual misidentification of sexes among sharks and rays, given the clear presence (male) or absence (female) of claspers. In this study, we present evidence of such discrepancies based on coverage differences across the entire heterogametic sex chromosome, which is less prone to potential false negatives compared to shorter molecular markers. This observation heightens the potential conflict between visual and genetic sex identification in white sharks.

We note here that there are criteria where both methodologies correctly detected sexual traits, yet the visual and genetic signals still contradicted each other. As pointed out by Devloo-Delva et al. [31], such a discrepancy could be caused by hermaphroditic (or intersexual) individuals. As intersexual individuals have been reported in multiple elasmobranch species (e.g., [32,33]), the white shark sequenced by Marra et al. may represent such a case. Alternatively, this discrepancy may be explained by the presence of rare sex-reversed individuals, as seen in some amphibians [34,35,36] and teleosts [37,38,39]. Although sex-reversed individuals have not been reported in elasmobranchs, the observed flexibility of the sexual traits in elasmobranch species (e.g., [32,33,40]) may suggest the possibility of complex or flexible sex determination systems that occasionally complicate sex identification in some species. Thorough investigations of the sexual phenotypes of white sharks, combined with sex-specific genetic markers or karyotypic examinations, can bring insights into the validity of these speculations. As the genomic data become more accessible for the study of shark and ray sex determination (e.g., [16,17]), future extensive surveys are likely to bring knowledge of both visual and genetic sexual cues and the mechanisms of sex determination.

Overall, this case study of a “female” individual that shows a strong male genetic signature provides an opportunity to investigate the techniques for accurate sex assignment or the exploration of alternative mechanisms. Accurate sex annotations for the genomic sequences are crucial for identifying those sex-specific sequences and the associated variations that are widely used for identifying sex chromosomes [18,41] and for designing sex-specific molecular markers (e.g., [31,42]). Insights from this case study highlight the importance of accurate and comprehensive documentation of sexual phenotypes from the individuals used in genomic assemblies.

It is reasonable to expand this observation and urge future material collection efforts for genome assemblies to provide as much information as possible, which can include but is not limited to detailed data on localities, body sizes, and developmental stages, along with clear photography. Providing such information can aid future efforts in disentangling potentially cryptic species or isolated populations segregated by morphological, behavioral, or chromosome structural differences (e.g., the Carolina hammerhead (Sphyrna gilberti) [43], Australian blacktip shark (Carcharhinus tilstoni) [44], and thorny skate (Amblyraja radiata) [45]). As genome assemblies are still relatively expensive and rare, investigations into matters such as comparative genomics and reproductive health can greatly benefit from each available genome assembly, and more so if the phenotypes of the source individual are thoroughly documented.

5. Conclusions

We demonstrated the efficacy of assigning genetic sexes to white shark individuals by mapping short-read sequences to a high-quality reference genome with identified sex chromosomes. This approach revealed a conflicting sex assignment for an individual previously used in a genome assembly. As the sex-determination mechanisms in the chondrichthyans are being explored, we believe it is essential for future genome assembly studies to document detailed sex information for the sourced samples, as it is currently unavailable for several assembled genomes. We extend this call for comprehensive sample information to cover the locality, body size, and developmental stage, as these details can maximize the utility of these genomes in chondrichthyan studies.