Selection of spermatazoa

Abstract

The present invention relates to the identification and / or selection of sperm, and in particular a means for identifying and / or selecting the fittest sperm for use in assisted reproductive technologies.

Description

[0001] Selection of Spermatazoa

[0002] Field of the Invention

[0003] The present invention relates to the identification and / or selection of sperm, and in particular a means for identifying and / or selecting the fittest sperm for use in, for example, assisted reproductive technologies (ART).

[0004] Background

[0005] In mammalian systems, males produce tens of millions of sperm in a single ejaculate, but hundreds reach the fallopian tube, and only one successfully fertilises an oocyte. Such a reduction in cell numbers suggests that sperm are exposed to a variety of obstacles and hence to natural selection prior to reaching the fertilisation site. Only those sperm capable of overcoming such challenges in vivo ultimately sire offspring. Such considerations invoke the heterogenous nature of an ejaculate. While most studies initially focused on sperm motility, shape and size classifications, subsequent research has elucidated the variation in sperm pools from within a single ejaculate in terms of energetics, capacitation and membrane composition (Keyser et al., 2022). Nonetheless, the most indisputable diversity among sibling sperm is the genetic variation resulting from segregation and recombination events during meiosis (Immler 2019; Bell et al., 2020). Such heterogeneity suggests that natural selection would favour sperm with the “best” phenotypic and genetic traits for fertilisation.

[0006] Evidence suggests that fertilisation success and offspring fitness is closely linked with the sperm phenotype and genotype. Studies in Danio rerio (zebrafish), the use of which has emerged as a reliable model for reproductive and fertility research due to its short cycle of reproductive period and the close degree of similarity of reproductive regulation systems with humans, have revealed that within-ejaculate sperm selection for longevity phenotypes affects offspring fitness. Longer-lived sperm sire offspring with higher developmental rate, reproductive performance and longevity (Alavioon et al., 2017, 2019). Such functional variation in phenotypes is directly linked with the variation in the underlying haploid genotype of the sperm (Alavioon et al., 2017). Other studies in the broadcast spawning ascidian Styela plicata and Atlantic salmon Salmo salar also linked the sperm longevity phenotype with higher hatching success and embryo survival and development (Immler et al., i 2014). These studies suggest that sperm phenotypic variation most likely results from the underlying haplotypic diversity; and that alleles might not be equally shared among sperm and as such are passed down the generations at different frequencies.

[0007] The suggestion that sperm haplotype influences its phenotype and offspring fitness challenges the current belief that the haploid spermatids are transcriptionally and translationally silenced (Grunewald et al., 2005; Baker, 2011). In fact, mounting evidence supports the idea of post-meiotic gene and protein expression. Studies in the house mouse Mus musculus and bulls Bos taurus have shown that sperm have de novo protein translation during capacitation in the female reproductive tract, and such activity is critical for sperm fertilisation ability (Gur & Breitbart 2006, 2008). Furthermore, in mice, genes such as Spami and Smpdi, Smoki have been shown to be actively transcribed and non-equally distributed among spermatids via cytoplasmatic bridges (Zheng et al., 2001; Butler et al., 2008; Veron et al., 2009). In fact, incomplete mRNA sharing appears to be much more widespread than previously thought across different mammals. Between 31 to 52% of haploid expressed genes have been shown to escape sharing and exhibit allelic biased expression in the house mouse Mus musculus, cattle Bos taurus, cynomolgus macaques Macacafascicularis and humans Homo sapiens (Bhutani et al., 2021).

[0008] By contrast, it is not currently understood if such allelic differences have phenotypic effects on mature sperm. If such assumptions are true, selection within an ejaculate might play a bigger role than previously thought, particularly in humans.

[0009] The quality of the sperm used in ART is thought to affect the success of the procedure. Sperm selection in an ART setting typically focuses on sperm motility, morphology and maturity. Current clinical options rely on the separation of motile from immotile sperm cells. Selection maybe based on observation of a sample under a microscope, with sperm selected based on observed morphology and motility. Alternatively or in addition, sperm may be selected based on membrane integrity, density and surface charge, whereby the sperm are phenotypically assessed through motility assays or differential gradients (Mcdowell et al., 2014). For example, sperm may be subjected to a swim-up assay, where a sample of the raw or washed ejaculate is placed in a tube with a suitable medium and incubated for a period of time; the top layer of the medium is then selected and sperm within it used for ART. As a further example, density gradient centrifugation (DGC) forces gametes to cross a gradient made of colloidal silicon and separates them based on their density. Whilst DGC is considered to yield sperm populations with higher motility, better morphology and maturity with respect to the raw ejaculate, evidence indicates that DGC may increase sperm DNA fragmentation, a parameter that negatively impacts reproductive outcomes after ART (Muratori et al. 2019).

[0010] Selection of sperm at the gametic level could provide an alternative or additional quality check. It is therefore desirable to understand which phenotypic and genotypic traits are under selection to be able to identify the ‘fittest’ sperm within an ejaculate. W02020 / 102565 is concerned with sorting sperm according to certain Geno-Informative Markers (GIMs). The disclosed approach does not use any link between sperm phenotype and sperm genotype for the sorting, but instead relies upon genetic markers.

[0011] It is amongst the objects of the present disclosure to address these problems and meet the needs discussed here.

[0012] Summary

[0013] The present disclosure provides evidence that functional heterogeneity in sperm is linked with underlying haploid genome diversity, by demonstrating allelic differences between sperm with different fitness levels. Genetic differences were found between sperm separated on the basis of a selection assay following a period of incubation or a performance based assay, with differences identified in the expression of associated downstream proteins (biomarkers) in sperm with improved fitness compared to less fit sperm. These findings provide basis for separating sperm on the basis of allelic differences, wherein sperm expressing a biomarker associated with an allele which is associated with improved fitness may be separated from sperm with a different allele at the same loci, which may, as a result, not express the biomarker, or may express a higher or lower level of the biomarker. This would allow sperm expressing the biomarker to be separated from a pool of sperm; or sperm exhibiting an increased or decreased level of expression of the biomarker relative to other sperm within a sperm pool to be separated from the pool of sperm; or sperm which do not express the biomarker to be separated from a pool of sperm.

[0014] Accordingly, in a first aspect there is provided a method for identifying a population of sperm with one or more alleles associated with improved fitness within a sample of sperm, comprising comparing the expression of one or more biomarkers associated with the one or more alleles to a threshold level for each of the one or more biomarkers; wherein the one or more biomarkers include Olfactomedin-4 (OLFM4) and / or Junction plakoglobin (CTNNG).

[0015] In some embodiments, at least one of the biomarkers is a negative biomarker, and sperm with improved fitness express below the threshold level of the biomarker. In some such embodiments, the population of sperm comprising one or more alleles associated with improved fitness in the sample express below the threshold level of OLFM4 and / or CTNNG.

[0016] In some embodiments, at least one of the biomarkers is a positive biomarker, and sperm with improved fitness express at least the threshold level of the biomarker.

[0017] In some embodiments, the one or more biomarkers further comprise one or more of phospholipase A2 (PLA2), beta-defensin 105 (DEFB105A), heterogeneous nuclear ribonucleoprotein K (hnRNP K or protein K), Midkine (MK), thyroxine-binding globulin (TBG), Afamin, Beta-microseminoprotein (MSP) and Alpha-i-acid glycoprotein 1 (A1AG1; also known as orosomucoid (ORM)).

[0018] In some embodiments, the population of sperm comprising one or more alleles associated with improved fitness in the sample express below the threshold level of PLA2, DEFB105A, hnRNP K and / or MK; and / or at least the threshold level of TBG, Afamin, MSP and / or A1AG1.

[0019] In some embodiments, the threshold reflects an expression level of at least a log2two-fold change relative to a control. In some embodiments, the threshold reflects an expression level of at least log2three-fold, at least log2four-fold, at least log2fivefold, or at least log2six-fold change relative to a control.

[0020] In some embodiments, the method further comprises determining the sperm population expressing below the threshold level of at least one of the biomarkers to be a first population; and the sperm population expressing at least the threshold level of the same biomarker(s) to be a second population. In some embodiments, the sperm sample is from a human or non-human mammal, optionally wherein the sperm sample is from a human. In some embodiments, the method does not comprise selecting sperm on the basis of carrying an X or Y chromosome.

[0021] According to a second aspect there is provided a method of selecting a population of sperm identified according to the first aspect, comprising: (a) separating the first population from the sperm sample, wherein the sperm in the first population constitutes the selected sperm; or (b) separating the second population from the sperm sample, wherein the second population constitutes the selected sperm.

[0022] According to a third aspect there is provided a method for assessing the reproductive potential of a sample of sperm, comprising: i) quantifying the number of sperm in a population selected using the method according to the second aspect; or ii) assessing the ratio of sperm in a population selected according to the second aspect, to the sperm within the sample which are not in the selected population.

[0023] In some embodiments, the methods according to the first, second or third aspect are for use in an assisted reproductive technology setting.

[0024] According to a fourth aspect, there is provided an enriched population of sperm comprising the one or more alleles associated with improved fitness, selected according to the method of the third aspect. In some such embodiments, the enriched population of sperm are for use in an assisted reproductive technology setting.

[0025] Brief Description of Figures

[0026] For a better understanding of the invention, and to show how embodiments of the same may be carried into effect, reference will now be made, by way of example, to the accompanying Figures, in which:

[0027] Figure 1: Genetic difference between unselected and selected sperm pools from nine male human donors: A: DMUKD1; B: DMUKD2; C: GADKM8; D: GADKM11; E: GADKM12; F: GADKM13; G: DMUKD4; H: DMUKD6a; I: DMUKD6B;

[0028] Figure 2: Distribution of gene type elements: average across all human donors and selected time points; Figure 3: Gene enrichment analysis of selected human sperm pools: A: sperm selected at 4 hours; B: sperm selected at 24 hours; C: sperm selected at 48 hours; Figure 4: Brain expressed genes with allelic deviations in the selected human sperm pools;

[0029] Figure 5: Cellular GO terms for the human sperm pools selected for improved fitness;

[0030] Figure 6: Dot plot showing sperm phenotypes under selection; (A) VCL (B) Chromatin maturity (C) Gross morphology and (D) DNA integrity. Each treatment group is labelled on the x-axis (R: Raw (n=35), W: Washed (n=35), SW: Swim-up (n=25), T4: Four-hour treatment (n=35) and T24: Twenty-four-hour treatment (n=i2)). Raw data is displayed in faded dots per each selection treatment group. The circular dots represent the mean values, while the error bars indicate the standard error;

[0031] Figure 7: A (Top left): Volcano plot displaying the proteins with a log fold change greater than 1 or below -1 for donor 1 (TMT run 3); B (Top right): Volcano plot displaying the proteins with a log fold change greater than 1 or below -1 for donor 1 (TMT run 2); C (Middle left): Volcano plot displaying the proteins with a log fold change greater than 1 or below -1 for donor 2 (TMT run 2). D (Middle right): Volcano plot displaying the proteins with a log fold change greater than 1 or below -1 for donor 1 (TMT run 1); E (Bottom left): Volcano plot displaying the proteins with a log fold change greater than 1 or below -1 for donor 2 (TMT run 1).

[0032] Detailed Description

[0033] Previous studies have suggested that sperm selected for phenotypes associated with greater longevity maybe genetically different (Alavioon et al., 2017). The evidence supporting this notion includes genome-wide genetic differences found between zebrafish sperm separated on the basis of a selection assay, where sperm were selected using a time based assay (where sperm selected following an incubation period were the selected population); and / or a performance based assay (where sperm that had swum through a viscous medium were the selected population) (Alavioon et al., 2017).

[0034] The haploid stage of spermatogenesis is the most transformative stage during male germline development, as morphological and biochemical changes occur during the passage from round spermatids to flagellated sperm cells. As a result of such changes, haploid spermatids are believed to be deprived of transcription (Baun, 1998). Nonetheless, a plethora of transcripts are expressed during the haploid stages, and these appear to not only be involved in transcriptional processes necessary for germline development, but also in multicellular processes such as organismal and nervous system development thus indicating that such haploid expressed transcripts are likely to play a key role in embryonic development and offspring fitness (Bhutani et al., 2021).

[0035] The evidence for haploid gene expression (Joseph and Kirkpatrick, 2004), the importance of selection during the haploid phase (reviewed in Immler, 2019) and variation among sperm within the same ejaculate (Immler et al., 2014; Bhutani et al. 2021) leads to a need to understand the contribution of mechanisms that maintain genetic variation within the haploid phase, such as additive genetic variance, balancing selection, and ploidally antagonistic selection.

[0036] The inventors thus undertook work to investigate the areas of phenotypic variation and resulting fitness. Human sperm samples selected using two selection assays were genetically compared (Example 1). The first selection assay was a performance based assay, where a sample of raw ejaculate was placed in the centre of a watchglass filled with methyl cellulose and the watchglass incubated for two hours. Sperm were then selected from the middle and edge of the watchglass. This methylcellulose performance-based assay is a known test for sperm motility and migration, providing an in vitro comparison to sperm penetration into cervical mucus. The test requires incubation of the sample with the methyl cellulose to allow sperm within the sample time to swim through the viscous methylcellulose. The second selection assay was a swim-up assay, where sperm were selected from a raw ejaculate using a swim-up assay without a period of incubation (To sample); or were selected using a swim-up assay following incubation of the raw ejaculate for a period of two (T2 sample), four (T4 sample), eight (T8 sample), 12 (T12 sample), twenty-four (T24 sample) or forty-eight hours (T48 sample). The findings revealed allelic divergence across the whole genome between sperm with improved fitness (as identified by the first and second assays; referred to in discussions of Example 1 as ‘selected sperm’) and less fit sperm (as identified by the first and second assays; referred to in discussions of Example 1 as ‘unselected sperm’). Whole genome sequencing data provided evidence that functional heterogeneity among human sperm is linked with its underlying haploid genome diversity, with remarkable allelic differences between selected sperm in comparison to unselected sperm within the same ejaculate. The loci deviating from the expected 50:50 allelic ratio with highest statistical significance were found in autosomal protein-coding regions and in long-noncoding RNAs (IncRNA). Selected sperm were found to be enriched in genes known to be involved in brain-specific functions, including neuron projections, synapse and cell junction organisation. Furthermore, selected sperm were not biased towards one sex. Such genetic differences can be magnified by selective effects on sperm phenotypic differences. These studies provided evidence that functional heterogeneity in sperm is linked with underlying haploid genome diversity, by demonstrating allelic differences between selected sperm (sperm with improved fitness) and unselected sperm (less fit sperm). These findings provide basis for genotypic sperm selection approaches for use in assisted reproductive technologies.

[0037] The inventors then investigated the effects of selection with and without an incubation period on sperm kinematics, morphology, chromatin and DNA integrity (Example 2). Key traits associated with fertilisation success were assessed in raw, washed and swim-up samples and compared to traits observed in sperm selected by swim-up following four or twenty four hours of incubation. Findings showed improvements in velocity, morphology and chromatin structure in selected sperm that had undergone an incubation period compared to unselected sperm and selected sperm that had not undergone an incubation period; but a decrease in DNA integrity in selected sperm that had undergone a period of incubation compared to selected sperm that had not undergone an incubation period.

[0038] Proteome analysis of human sperm was then undertaken (Example 3). Sperm were selected using two assays. Firstly, a performance-based assay, where a washed sample of raw ejaculate was placed in the centre of a watchglass filled with methylcellulose and sperm were selected from the middle and edge of the watchglass after four hours. Sperm at the edge of the watchglass were considered to be sperm with improved fitness; sperm collected from the centre of the watchglass were considered to be less fit sperm. Secondly, a swim-up assay, where sperm were selected from a washed raw ejaculate sample using a swim-up assay without a period of incubation (To sample); or were selected using a swim-up assay following incubation of the washed raw ejaculate sample for a period of four hours (T4 sample). Selected sperm in the T4 sample (i.e. which had undergone a period of incubation) were considered to be sperm with improved fitness; sperm in the To sample (i.e. which had not undergone a period of incubation) were considered to be less fit sperm. Proteome analysis was carried out on the sperm with improved fitness and the less fit sperm, and genes associated with differences between the two groups, and their downstream proteins (biomarkers) were identified. The expression of these biomarkers differed between the sperm with improved fitness and the less fit sperm. The inventors then selected 2 of the cell surface biomarkers identified in Example 3, OLFM4 and CTNNG, for validation using fluropho reconjugated antibodies (Example 4). The results found significantly more expression of OLFM4 and CTNNG in less fit sperm, thus demonstrating the use of OLFM4 and CTNNG as negative biomarkers of sperm fitness.

[0039] These findings provide basis for separating sperm on the basis of allelic differences, wherein sperm with an allele which is associated with improved fitness maybe separated from sperm with a different allele at the same locus.

[0040] The present invention thus resides in the differential expression of the identified biomarkers within a sperm population. It should be noted that, while CTNNG and OLFM4 were demonstrated to be negative biomarkers for sperm fitness (i.e. high levels of expression are associated with less fit sperm). Other biomarkers identified in Example 3 are positive biomarkers (i.e. high levels of expression are associated with sperm with improved fitness). The skilled person would have no difficulty establishing which of the biomarkers identified in Example 3 are positive biomarkers and which are negative biomarkers.

[0041] Accordingly, in a first aspect there is provided a method for identifying a population of sperm with one or more alleles associated with improved fitness within a sample of sperm, comprising comparing the expression of one or more biomarkers associated with the one or more alleles to a threshold level for each of the one or more biomarkers; wherein the one or more biomarkers include OLFM4 and / or CTNNG.

[0042] The term ‘improved fitness’ as used herein refers to enhanced performance as assessed in either a selection assay following a period of incubation, where the sperm considered to have enhanced performance (and therefore improved fitness) are the sperm that perform best in the assay (for example, the sperm in the top layer of a swim-up assay which is conducted following incubation of the sperm sample for a period of time (for example, 2 or 4 hours, and typically, 4 hours) are considered sperm with improved fitness; sperm which are not in the top layer are considered less fit sperm); or a performance assay, where sperm collected from the outer edge of a watchglass filled with methylcellulose after a period of time (for example 2 or 4 hours, and typically, 4 hours after commencing the assay) are considered sperm with improved fitness; sperm which are not in the outer edge are considered less fit sperm. The sperm sample maybe a raw ejaculate, or may have undergone processing and / or handling in preparation for use in a selection assay (for example washing of the raw ejaculate in human tubal fluid and condensing).

[0043] The methods of the present disclosure recognise differences in the level of biomarker expression within a sperm sample. It is known that the parameters and quality of a sperm sample vary significantly between individuals. In keeping with this, expression levels of the one or more biomarkers may vary significantly between individuals. As such, the differential expression of one or more biomarkers between sperm with improved fitness and other sperm in the sample is individual to each sample. For example, in an embodiment wherein the biomarker is OLFM4, and sperm with improved fitness express low levels of OLFM4, a sample from one individual may contain a high number of sperm which strongly express OLFM4 and a small number of sperm which weakly express OLFM4; a sample from another individual may contain a few sperm which strongly express OLFM4 with the remaining sperm in the sample weakly expressing OLFM4; a sample from a further individual may contain a high number of sperm which very weakly express OLFM4, with only a small number of sperm which moderately express OLFM4. It is desirable, using the methods of the present disclosure, to identify, detect and separate the sperm with the lowest levels of OLFM4 within each of these samples from the other sperm within that sample.

[0044] A skilled reader would understand how this could be achieved. For example, a threshold level of expression of the biomarker may be determined for each sample. If the biomarker is a positive biomarker, expression of the marker at or above the threshold value is associated with sperm with increased fitness, which therefore have reasonable or good reproductive potential. If the marker is a negative biomarker, expression of the biomarker below the threshold value is associated with sperm with increased fitness which therefore have reasonable or good reproductive potential.

[0045] In some preferred embodiments, the threshold level is set to reflect the minimum signal at which sperm cells expressing the biomarker(s) in question can be reproducibly separated from the rest of the population.

[0046] In some embodiments, the threshold value may be determined quantitatively. For example, levels of expression of the one or more biomarkers may be compared to the lowest or highest level (as appropriate for that biomarker) of expression of that biomarker within the sample; or to a control. For example, when using fluorimetry, an undyed sperm cell from the sample to be assessed may be included to act as a control. In some embodiments, the use of such an ‘internal’ control is preferred, as the threshold level can vary, not just between samples, but also depending upon the means used to detect expression. For example, fluorescence levels vary depending upon the fluorop hore and / or means of detection used, as a skilled person would understand.

[0047] In some embodiments, sperm may be detected which express at least a log2two-fold change in expression level of the one or more biomarkers compared to sperm expressing the lowest or highest level of the biomarker (as appropriate for that biomarker); or to a control; and the threshold level set accordingly. In some embodiments, the threshold level maybe set based on at least at least log2three-fold, at least log2four-fold, at least log2five-fold, or at least log2six-fold change in expression compared to sperm expressing the lowest or highest level of the biomarker (as appropriate for that biomarker); or to a control.

[0048] For non-flow cytometry detection means, a fluorescence-based threshold can be translated into an equivalent metric based on surface biomarker density or binding affinity: a fluorescence signal corresponds to an estimated number of bound probe molecules per cell, which can be converted into a surface density (for example molecules per pm2of plasma membrane) or total biomarker copy number per cell. Thus, in non-flow separation platforms (such as magnetic, microfluidic, or affinitybased capture), the same discriminatory threshold can be applied by calibrating the capture reagent or detection modality to recognize only those cells with biomarker expression above the equivalent of the threshold determined by fluorescence. In this way, the flow cytometry intensity value serves as a quantitative benchmark that can be directly transposed into a functional threshold for alternative separation technologies.

[0049] In alternative embodiments, the threshold value may be determined manually, for example following observation based on the visible level of expression (for example, based on the level of fluorescence).

[0050] In some embodiments, the one or more biomarkers are as listed in Table A. In preferred embodiments, the one or more biomarkers are as listed in Table B. In some embodiments, the one or more biomarkers comprise one or more of OLFM4, CTNNG, phospholipase A2 (PLA2), beta-defensin 105 (DEFB105A), heterogeneous nuclear ribonucleoprotein K (hnRNP K or protein K), Midkine (MK), thyroxine- binding globulin (TBG), Afamin, Beta-microseminoprotein (MSP) and Alpha-i-acid glycoprotein 1 (A1AG1; also known as orosomucoid (ORM)). In some embodiments, biomarkers OLFM4, CTNNG, PLA2, DEFB105A, hnRNP K, MK, TBG, Afamin, MSP, and / or A1AG1 are considered, of the biomarkers listed in Tables A and / or B, preferred for use for identifying sperm for use in an ART setting. These markers demonstrated significant differential expression in sperm selected following a period of incubation according to Example 3 (T4 and T24 sperm) in comparison to unselected sperm or sperm selected without an incubation period (To sperm).

[0051] These biomarkers are also cell-surface expressed. In some embodiments the biomarkers comprise OLFM4 and / or CTNNG. In some embodiments the biomarkers comprise OLFM4. In some embodiments the biomarkers comprise CTNNG. In some embodiments the biomarkers are OLFM4 and / or CTNNG. In some embodiments, the biomarker is OLFM4. In some embodiments the biomarker is CTNNG.

[0052] In embodiments according to the first aspect, identifying sperm in the sample which express one or more biomarkers or a threshold level of one or more biomarkers may be achieved using any suitable means, such as any suitable cell labelling means. In preferred embodiments, one or more of the biomarkers are a surface-expressed protein. In some such embodiments, identification may comprise contacting the sample with a label (such as a molecule or protein) that binds to the biomarker. For example, the sample may be contacted with an antibody molecule that binds to the biomarker. The antibody may comprise a label, such as a fluorescent, colour, magnetic or size label, or other means of identification. For example, the antibody maybe conjugated to biotin, which can complex with a fluorescence-carrying label; or to a label of a certain size. Detection of labelled sperm or sperm with a threshold level of the label may be achieved using known means, such as fluorescence or size detection. For example, a cell sorting means, such as those known in the art, can be calibrated to detect sperm with a sufficient level of the label to reflect the desired expression level. For example, an antibody molecule maybe used that binds to one or more of the biomarkers and has a fluorescent, colour, size or magnetic label. The selected detection means is capable of quantifying the level of fluorescence, colouration, size or magnetic signal.

[0053] In some embodiments, the method according to the first aspect further comprises determining the sperm population expressing below the threshold level of at least one of the biomarkers to be a first population; and the sperm population expressing at least the threshold level of the same biomarker(s) to be a second population.

[0054] The sperm sample maybe from a human or animal. In some embodiments, the sperm sample is from a human or a non-human mammal, or a fish. The sperm sample may be fresh (i.e. an ejaculate that has not been frozen or cryopreserved). Alternatively, the sperm sample may be a frozen or cryopreserved ejaculate that has been thawed. In some preferred embodiments, the method does not comprise selecting sperm on the basis of carrying an X or Y chromosome.

[0055] In a second aspect there is provided a method of selecting a population of sperm identified according to the first aspect, the method comprising separating a first, target population (‘the selected sperm’) from the heterogenous mixture (the sperm sample); or separating a second target population (‘the selected sperm’) from the heterogenous mixture (the sperm sample).

[0056] Labelled sperm, or sperm which express at least a threshold level of the label may be separated from unlabelled sperm, or sperm which express below a threshold level of the label using any suitable cell sorting technique. For example, the sample may be contacted with an antibody molecule that binds to the biomarker. The antibody may comprise a label, such as a fluorescent, colour, magnetic or size label, or other means of identification. The antibody may, as an example, be conjugated to biotin, which can complex with a fluorescence-carrying label; or to a label of a certain size. A cell sorting means, such as those known in the art, can be calibrated to detect sperm with a sufficient level of the label to reflect the desired expression level. For example, an antibody molecule may be used that binds to one or more of the biomarkers and has a fluorescent, colour, size or magnetic label. The selected separation means is capable of quantifying the level of fluorescence, colouration, size or magnetic signal and separating sperm with an amount of label that reflects the desired expression level.

[0057] In some embodiments, the separated sperm comprise, substantially comprise or consist of sperm comprising one or more alleles associated with improved fitness. Alternatively, the separated sperm may comprise, substantially comprise or consist of less-fit sperm (i.e. sperm which do not comprise the one or more alleles associated with improved fitness).

[0058] The present disclosure further provides a means of assessing the reproductive potential, as based on fitness, of a sample of sperm. The term ‘reproductive potential’ as used herein means the capability to produce offspring.

[0059] Accordingly, in a third aspect there is provided a method for assessing the reproductive potential of a sample of sperm, the method comprising quantifying the number of sperm in a population selected using a method according to the second aspect; or assessing the ratio of sperm in the sample which are in a population selected using a method according to the second aspect, to the sperm within the sample which are not in the target population.

[0060] The quantification and / or assessment may be qualitative or quantitative. In some embodiments wherein the method of the first or second aspect involves labelling of a target population, the quantification and / or assessment maybe carried out by comparing the labelled population to the unlabelled population, for example using a spectrophotometer.

[0061] The method according to the second aspect allows preparation of a population of sperm having a greater probability than random selection of comprising one or more alleles associated with improved fitness; and / or having the same or a greater probability of comprising one or more alleles associated with improved fitness than sperm selected using selection assays typically used in the field of ART. Accordingly, in a fourth aspect there is provided an enriched population of sperm comprising the one or more alleles associated with improved fitness, selected according to the second aspect.

[0062] In some embodiments, the sperm in the enriched population comprise, substantially comprise or consist of sperm having greater than 50% likelihood of comprising one or more alleles associated with improved fitness, or greater than about 55%, greater than about 60%, greater than about 70%, greater than about 75%, greater than about 80%, greater than about 85%, greater than about 90%, greater than about 91%, greater than about 92%, greater than about 93%, greater than about 94%, greater than about 95%, greater than about 96%, greater than about 97%, greater than about 98%, greater than about 99% or greater than about 99.5% likelihood of comprising one or more alleles associated with improved fitness. In some embodiments, the sperm in the enriched population comprise, substantially comprise or consist of sperm having about 100% likelihood of comprising one or more alleles associated with improved fitness. In some embodiments, the enriched population maybe used in an ART setting.

[0063] There is further disclosed a method for identifying sperm with one or more alleles associated with improved fitness within a sample of sperm, comprising comparing the expression of one or more biomarkers associated with the one or more alleles to a pre-determined standard for each of the one or more biomarkers, wherein a significant difference in expression of the one or more biomarkers in the sample as compared to the pre-determined standard for each of the one or more biomarkers indicates that the sperm has one or more alleles associated with improved fitness; wherein the one or more biomarkers include OLFM4 and / or CTNNG. There is further disclosed a method for identifying a population of sperm comprising one or more alleles associated with improved fitness within a sample of sperm, wherein sperm within the sample differentially express one or more biomarkers associated with the one or more alleles; the method comprising providing a sperm sample; and detecting sperm in the sample which express below a threshold level of the one or more biomarkers; wherein the one or more biomarkers include OLFM4 and / or CTNNG. In some embodiments, the one or more biomarkers further comprise one or more of PLA2, DEFB105A, hnRNP K, MK, TBG, Afamin, MSP and A1AG1. In some embodiments the biomarkers are OLFM4 and / or CTNNG. In some embodiments, the biomarker is OLFM4. In some embodiments the biomarkers is CTNNG.

[0064] The methods and enriched population of sperm according to the abovementioned aspects may be used in an assisted reproductive technology setting.

[0065] Advantages

[0066] The present disclosure provides basis for developing genotypic selection options for sperm with improved fitness from within a sperm sample. As will be appreciated, and as evidenced by the data provided in the Examples from different human volunteers, the parameters and quality of sperm within a sample (sperm count, sperm concentration, motility, morphology, ejaculate volume) can vary enormously between individuals, as well as within the sample. Currently, sperm may be selected in a clinic for use in ART based upon an assay, such as a swim-up assay, although this may not be the case or not preferred, due to the potential duress it places on the sperm. Alternatively, or in addition, selection may be based on observation of the sperm sample under a microscope, with sperm that are deemed to be the ‘fittest’ selected based on criteria such as observed morphology and motility. The threshold for selection is relative, as it will depend upon the quality of the sperm within the sample, and typically it is desirable to select the ‘fittest’ sperm in the sample, even when the overall quality of the sample is poor.

[0067] The present disclosure provides a method of identifying and selecting sperm with improved fitness within a sample based on expression of one or more biomarkers. The identified biomarkers are cell-surface expressed, offering the advantage of ease of detection. The development of assays based on the identified biomarkers would offer the possibility for a faster and more reliable sperm selection process, which would avoid the need for motility assays, thereby avoiding any associated DNA fragmentation and / or damage, which in turn is likely to improve ART outcomes, and avoids reliance on selection of sperm based on observable characteristics such as motility and gross morphology, which may be subjective, and may be difficult to assess and / or quantify.

[0068] In addition, genotypic sperm selection could allow for a cheaper screening tool against suboptimal known alleles, since selection at the haploid level is more efficient against recessive alleles (Kondrashov and Crowt, 1991). It could also provide a less invasive intervention to current pre-implantation genetic testing options, where multiple embryos are biopsied and genetically tested for monogenic diseases or chromosomal abnormalities.

[0069] The robust nature of the biomarker data obtained in the present disclosure is illustrated by the use of sperm samples from different donors, and the use of two different assays ((i) a methylcellulose performance based assay; and (ii) a swim-up following a period of incubation) to select the sperm, which nevertheless gave rise to a similar pattern of biomarker expression in selected sperm (Example 3). This is important when considering translation of the present findings to an ART setting, where variations in sperm selection may occur between clinics, and where samples from many different donors are handled. The findings of Example 3 demonstrate that the identified biomarkers are consistent across sperm from several different donors, selected using the two different assays. Two of the identified biomarkers were then validated, demonstrating their use as negative biomarkers of sperm fitness (Example 4).

[0070] It is of further interest that at least some of the identified biomarkers have been associated with certain disease states, including cancer, autism, ADHD and neurodevelopmental diseases. It is possible that selection of sperm on the basis of decreased levels of expression of such biomarkers could impact on the disease susceptibility of an individual resulting from a successful pregnancy using the sperm. Such a suggestion finds support in studies in fish linking sperm longevity phenotype to improved embryo survival, as discussed in the background section provided here.

[0071] In addition, the findings of Example 2 demonstrate that motile sperm selected by swim-up after a period of incubation (T4 and T24) showed improvements in traits known to be associated with fertilisation success and embryo development (improved velocity, morphology, chromatin integrity and DNA integrity; Figures 6A- D) when compared to sperm in unselected sperm pools; and improvements in velocity, morphology and chromatin integrity when compared to sperm from a raw ejaculate selected using swim-up (To sperm) (Figures 6A, B and C). As such, selection of sperm using methods according to the present disclosure for use in an ART setting could positively affect fertilisation success, reproductive success (cumulative live birth rate per intention to treat; pregnancy per transfer; live birth rate; cumulative life birth rate per started cycle or per transfer) and / or embryo development. Furthermore, the findings of Example 2 demonstrate that DNA integrity decreases in motile sperm selected by swim-up after a period of incubation (T4 and T24) compared to sperm from a raw ejaculate selected using swim-up (To sperm) (Figure 6D). These results indicate that, whilst a period of incubation is beneficial for selection on the basis of certain advantageous traits (velocity, morphology, chromatin maturity), it has a detrimental effect upon DNA integrity, which could negatively affect fertilisation success, reproductive success and / or embryo development. The use of the biomarkers identified in the present disclosure would allow selection of sperm with known beneficial traits (velocity, morphology, chromatin maturity) from a sample, whilst avoiding the need to subject the sample to an incubation period (with its associated detrimental effects upon DNA integrity) in order to identify these sperm.

[0072] Further to this, it has been shown in zebrafish that within-ejaculate sperm selection for longevity phenotypes affects offspring fitness, with longer-lived sperm siring offspring with higher developmental rate, reproductive performance and longevity (Alavioon et al., 2017, 2019). Whilst it is not possible to draw a direct parallel to other species, and without being bound by a particular theory, it is possible that use of sperm selected by methods according to the present disclosure in an ART setting could confer similar benefits on resultant offspring.

[0073] The following Examples are provided to illustrate embodiments of the present invention and should not be construed as limiting thereof.

[0074] Example 1: Haploid selection causing allele frequency divergence among human sperm from within an ejaculate

[0075] Samples

[0076] Ejaculates from nine self-reported healthy volunteers with sperm concentration >15 mil / ml were collected near the laboratory and allowed to liquefy for 30 minutes at 37°C. Prior to the selection assays (discussed below), 2ooul of the raw ejaculate from each donor was stored at -8o°C to be used as a reference for allele frequency estimation. Sperm were then selected using two different assays, as detailed below. In vitro sperm selection assay i

[0077] Sperm were selected from the raw ejaculate from four donors (GADKM8, GADKMn, GADKM12 and GADKM13) using a first sperm selection assay: A 10 cm-diameter watch glass was filled with methylcellulose up to a 4cm radius. 300-400 ul of raw ejaculate was placed in the centre of the glass (the ejaculate sank and filled a circle of 0.5 cm in the centre of the glass). Five watch glasses were used per donor to enable sufficient sperm to be retrieved for subsequent sequencing. The watch glasses were incubated for 2 hours at 37°C. A 4 mL sample was then collected from the outer circle of each watch glass, at a radius of 3-3.5 cm, 2.5-3 cm or 2-2.5 cm (depending on the distance travelled by the sperm), and placed in a tube (edge sample). An additional 2 mL of medium was collected from the top of the centre of each watch glass at a distance of at least 0.5 cm from the bottom of the watch glass to avoid contamination with dead and somatic cells and placed in a tube (centre sample). The samples from the edge and centre respectively were pooled for each donor, to give one edge sample and one centre sample per donor. The samples were centrifuged and the sperm pellet was collected from the bottom. Sperm pellets were stored at - 800C for DNA extraction.

[0078] In vitro sperm selection assay 2

[0079] Sperm were selected from the ejaculate of five donors, (DMUKDi, DMUKD2, DMUKD4, DMUKD6a and DMUKD6B) using a second selection assay. 1 ml raw ejaculate was placed in each of five 15 ml centrifuge tubes at a 1:1 ratio with human tubal fluid (HTF+) medium containing (in mM): 72.8 NaCl, 4.69 KC1, 0.2 MgSO4, 0.37 KH2PO4, 2.04 CaC12, 0.33 sodium pyruvate, 21.4 sodium lactate, 2.78 glucose, 21 HEPES, 25 NaHCOs, adjusted to pH 7.4 with NaOH.

[0080] One tube was used for an immediate swim-up assay: the tube was flicked to homogenise the sample and iml was transferred to the bottom of a new 15 ml centrifuge tube containing 4ml of HTF+ medium and left to swim up at 37°C for 60 min in 5% CO2. The top-layer (3.5 ml) was then collected using a pipette (To sample).

[0081] The other four tubes were incubated at 37°C with 5% CO2 for 2 (T2), 4 (T4), 8 (T8), 12 (T12), 24 (T24) or 48 (T48) hours. Following each time point (To to T48), a swim-up assay was performed as detailed above and the top-layer (3.5 ml) in each case collected using a pipette.

[0082] The collected samples were centrifuged at ooxg for 10-20 min at room temperature (RT) prior to resuspension in 2 ml HTF. After assessing sperm concentrations of samples using image cytometry, the cells were pelleted and kept at -8o°C until further use.

[0083] DNA extraction

[0084] All samples from both assays were standardised to 2M cells / ml and DNA extraction was performed using the standard phenol:chloroform:isoamyl alcohol protocol. Briefly, the sperm pellets were lysed overnight at 55°C in 250 pl of proteinase K buffer (lomM Tris-HCL, womM NaCl, 25mM EDTA, 1% SDS) and 2.5 pl 1M DTT and 5 pl proteinase K (20 000 U). After overnight incubation, the samples were centrifuged at 8000g for 3 minutes after which 250 pl saturated phenol was added to the samples and shaken vigorously by hand for 7 minutes before centrifuging at 13000g for 5 minutes. 225 pl of the aqueous phase was then placed in a new 2ml microcentrifuge tube to which 225UI of phenol-chloroform was added. The samples were again hand shaken for 7 minutes and then centrifuged at 13000g for 5 minutes. 2iopl of the aqueous phase was placed in a fresh 2ml microcentrifuge tube to which 2iopl of chloroform was added. This was again shaken for 7 minutes and centrifuged at 13000g for 5 minutes. 200 pl of the top aqueous phase was transferred to a fresh 2 ml microcentrifuge to which 20 ul of 3 M sodium acetate and 500 pl 100% cold ethanol was added. Upon gentle inversion of the samples, these were precipitated for at least ih at -20°C. Samples were then centrifuged at 13 000 g for 15 minutes and the pellets were resuspended in 500 pl of 70% ethanol. Upon further centrifugation at 13 000 g for 15 minutes the pellets were dried and resuspended in 6oul of Tris- EDTA buffer before being stored at -8o°C until further processing.

[0085] Sperm pool sequencing

[0086] The extracted genomic DNA from each sample was prepared using the PCR-free library (KAPA HyperPrep kit; Roche) for a standard insert size (400-500 bp). Between 5oong-ipg of genomic DNA in a volume 55pl was sheared to 550bp using the Covaris Sonicator (Covaris). The resulting DNA was then size selected utilising KAPA Pure beads, selecting for fragments of 450bp-650bp in size. The ends of the DNA were repaired; 3' to 5' exonuclease activity removed the 3' overhangs and the polymerase activity filled in the 5' overhangs creating blunt ends. A single ‘A’ nucleotide was added to the 3’ ends of the blunt fragments to allow for the ligation of barcoded adapters at a concentration of 6pM prior to a bead clean up using KAPA Pure Beads. The quality of the resulting libraries was determined using a High Sensitivity DNA Kit from Agilent Technologies (5067-4626) and the concentration measured with a High Sensitivity Qubit assay from ThermoFisher (Q32854) finally q-PCR was carried out to allow for accurate pooling prior to sequencing.

[0087] Whole-genome paired-end sequencing libraries were sequenced on an Illumina NovaSeq 6000 platform to a mean coverage of ~25x per lane in 4 lanes per sample at a mean of 1.3 million reads per pooled sample.

[0088] Genome data analyses Trimmomatic 0.39 was used for adapter removal (present in TruSeq3-PE-2.fa) and quality trimming, followed by alignment of individual lanes to the Homo sapiens GRCI138 reference assembly with Burrows-Wheller aligner (BWA) 0.7.13 bwa-mem and removal of marked duplicate reads with Picard 2.1.1. Afterwards, SAMtools flagstat 1.3 was used on each bam file to extract the mapping statistics and SAMtools sort was used to sort each file. Coverage for each sperm pool was determined for autosomes and sex chromosomes using SAMtools coverage option. The variant detection of the alignments of reads from the centre / To samples followed the best practice workflow recommended by GATK. In brief, GATK 4.9.1 Haplotype Caller was used to call the variants and filter the heterozygous paternal sites. Single nucleotide polymorphisms (SNPs) were filtered using SelectVariant and the SNP filter expression was set as QD < 2.0 && FS > 60.0 && MQ < 40.0 && SOR > 3.0 && MQRankSum < -12.5 && ReadPosRankSum < - 8.0 && QUAL < 30. Alignment of the reads from centre / To samples and edge / T4 samples were used to generate base count profiles using pileup2pro2. SAMtools 1.9 option -q3O, -Q30 and diooooo was used to exclude sites with low quality, allowing for sites with deep coverage.

[0089] Next, the heterozygous sites and base count profiles of the centre / To samples and edge / Tq samples were used for likelihood ratio tests using the scrip hetPoolLikelihoods.pl (available at https: / / github.com / douglasgscofield / gameteUtils) based on Lynch et al (2014). The most likely allelic differences between the centre / To and edge / T4 samples were then identified by the largest likelihood ratios. Briefly, the major and minor alleles were assigned according to counts within the centre / To samples. Sites with a minimum of 32X and a maximum of 2*total mean coverage (mean coverage was 206.69, 200.68, 176.31, 174.9, 261.17, 221.03, 271.5, 239.1023 and 211.443 for donor GADKM8, GADKM11, GADKM12, GADKM13, DMUKD1 and DMUKD2, DMUKD4, DMUKD6a and DMUKD6B respectively) were excluded. The error rate was estimated based on the number of putative errors reads / number of total reads (Lynch et al. Eq 3a). Then, OM and Om were calculated for the probability of a random read being recorded as major and minor allele, respectively (Lynch et al. Eq la and ib). Next, p was used to estimate the major allele frequency in the pool accounting for the putatively non-erroneous and erroneous reads that represent the major allele type (Lynch et al. Eq 3b). The log likelihood of the data following the model as well as data under the assumption of monomorphism was then computed following Lynch et al. Eq 4a and 4b, respectively. Finally, the likelihood ratio test statistic was computed following a Chi-squared distribution with 1 degree of freedom (LRT_Maf in hetPoolLikelihoods.pl). To identify statistically significant loci between centre / To and edge / Tq samples, the threefold for genome-wide significance was set higher than the 99thquantile; critical values were 17.08, 11.91, 15.36, 11.60, 11.37, 11.30, 11.67, 11-37 and 11.6 for donor GADKM8, GADKM11, GADKM12, GADKM13, DMUKDi, DMUKD2, DMUKD4, DMUKD6a and DMUKD6B respectively.

[0090] To investigate which genetic elements reside on the statistically significant loci, the loci of interest were overlapped with the annotated reference genome using the fOverlaps function from TREGELvo.0.1.1 package (Berres et al. unpublished), hereby the significant loci were provided as “query” and the annotated genome as “subject”. In quest of elucidating the biological significance of the genetic differences resulting from the comparisons between centre / To and edge / Tq samples, the resulting overlaps were then used for enrichment analysis using ShinyGO vo.75 (Ge et al., 2020).

[0091] Results

[0092] Sperm selection

[0093] The assays used selected sperm on the basis of fitness, eliminating shorter-lived, less motile and / or immotile sperm cells. In fact, a dramatic reduction in cell numbers was observed at 4 hours of selection. At 24 and 48 hours there was a general decrease in cell numbers compared to 4 hours. These results indicated that there is a great variation between donors, as a consequence of their heterogenous ejaculate and that only a modest proportion of sperm from within the same ejaculate have the ability to survive longer periods of time (Table 1).

[0094] Table 1: Sperm concentration of donated samples from 5 donors

[0095] Linking improved fitness phenotype with genotype

[0096] A statistical method similar to the one developed in Alavioon et al. (2017) was used to estimate allele frequency in a population of sperm based on likelihood ratio tests. The heterozygous sites from each raw ejaculate were used to estimate allelic deviations in the centre / To samples. A raw ejaculate is known to contain a heterogeneous population of cells which adhere to Mendelian allele frequency expectation, whereby the probability of each gamete inheriting any allele type is 50% (Carioscia et al. 2021). Therefore, any allele frequency deviation from this ratio in any of the edge / Tq samples would strictly be the result of haploid selection. Although considerable variation was observed between donors and sperm pools, it was found that less than 0.05% of the loci compared between raw ejaculates and edge / Tq samples showed statistically significant allelic deviations in all donors across the different time points (Table 2). Each such site indicates the location of the deviation from the expected 50:50 allelic ratio. Most deviations across donors and samples occurred in autosomal regions and the strongest signals were present near the centromeres in the samples selected at 2 h from donors GADK8, GADK11, GADK12 and GADK13 (Figure 1). Table 2: Total number of loci analysed in likelihood ratio tests and significant loci resulting from the analysis Phenotypic sperm selection and sperm sex ratios

[0097] The genomic coverage approach was used to identify sex chromosomes in the pooled centre / To and edge / T4 samples (see Figure 2). The pooled results were grouped by either centre / To or edge / T4 samples and performed a Student’s t-test for each group. No evidence for X or Y chromosome skews in the centre / To sperm pools (p- value=o.2) was found, as expected. In addition, no differences in sex chromosome coverage were found in the edge / T4 sperm pools selected at 2 hours (p-value=o.46), 4 hours (p-value=o.4i), 24 hours (p-value=O-48) or 48 hours (p-value=0.44), indicating that the 50:50 sex ratios were maintained in the sperm pools selected for improved fitness.

[0098] Selection on sperm performance in coding and non-coding genomic regions Following the identification of statistically significant genomic loci, where alleles deviate from the expected Mendelian ratios between sperm pools, the genes which might be present at those exact locations in the genome was investigated. TREGEL was used to identify the genetic elements present in the datasets generated from the likelihood ratio tests. A series of genetic elements were identified including unprocessed pseudogenes, transcribed unprocessed pseudogenes, TEC, processed pseudogenes, unitary pseudogenes, polymorphic pseudogenes, IncRNA, snoRNA and protein coding genes (Figure 3). Interestingly, the predominant elements in the datasets across all donors and sperm pools (irrespective of the selection time) were protein coding genes and IncRNA.

[0099] Sperm performance is linked with cellular genotype

[0100] Given the abundance of genetic elements present in the datasets across donors and sperm pools, enrichment analysis was performed to identify potential pathways contributing to the performance of the edge / T4 sperm populations. ShinyGo (Ge et al., 2020) was used to perform gene enrichment analysis on the different selected sperm pools datasets using a cut off FDR p-value of 0.05 (Figure 4 and Figure 5).

[0101] The top cellular function GO terms across time points and donors were enriched in post synaptic membrane, neuron projections and synapse, particularly in the samples selected from 4 hours onwards (Figures 4 and 5).

[0102] Notably, five genes showed significant divergence between sperm pools across all donors: DLG2, PTPRD, CSMD1, CDH18 and NRXN1. DLG2, PTPRD, CSMD1, CDH18 are all well-known tumor suppressor genes, whereas CSMD1, CDH18 and NRXNi are associated with developmental disorders including neurodevelopmental disorders, as well as senescence, schizophrenia and ASD.

[0103] With regards to the genotypes under potential selection underpinning the sperm fitness phenotypes, it is shown that sperm with improved fitness are enriched in 50.6-81.7% of genes which also show strong expression in the brain (Figure 1) involved in neuron projections, synapse and cell junction organisation (Figure 4).

[0104] Discussion

[0105] The whole genome sequencing data of a total of 25 sperm pools from nine individuals presented here provides evidence that the functional heterogeneity among human sperm is linked with its underlying haploid genome diversity. Despite considerable variation between samples and assays, remarkable allelic differences between sperm with improved fitness and shorter-lived or immotile cells within the same ejaculate were found in all comparisons. The loci deviating from the expected 50:50 allelic ratio with highest statistical significance were found in autosomal protein-coding genes and one-third fell in long-noncoding RNAs. Such genetic differences can be magnified by selective effects on sperm phenotypic differences. This evidence provides basis for the development of genotypic sperm selection approaches for use in assisted reproductive technologies.

[0106] In more detail: the number of motile sperm pertaining the ability to penetrate the medium in swim-up assays declined by more than three-fold after 4 hours compared to To samples from the same individual. In humans, in vivo, sperm are deposited in the anterior vagina during ejaculation. The sperm leave the seminal plasma within minutes and swim up the cervix (Suarez and Pacey, 2006). Whilst little is known about the longevity of sperm in the female reproductive tract, reports suggest that, although in minimal amounts, motile sperm have been recovered from the cervix or uterus 50 hours and even 125 hours after insemination (Rubenstein et al., 1951; Moyer et al., 1970; Gould et al., 1984). In vivo, sperm are under strong natural selective pressure to maximise the potential fertilisation processes.

[0107] The present results show that in vitro using HTF+ medium motile sperm can be recovered even after 48 hours of culture, although fewer cells could be obtained at later incubation time points. Contributing further to the knowledge of previous studies, it has been possible to provide far greater details about the cellular and molecular function of those recovered longer-lived sperm.

[0108] It has been previously suggested that sperm selected for improved fitness phenotypes are genetically different (Alavioon et al., 2017). The key evidence supporting this notion includes genome-wide genetic differences between longer and shorter-lived zebrafish sperm.

[0109] The presented investigation revealed allelic divergence across the whole genome between centre / To and edge / T4 human sperm. Whilst two different selection assays were performed at different time points, strong genotypic evidence supporting phenotypic findings were found in both. To mimic in vivo conditions, two different assays were developed, and cells selected at 2, 4, 8, 12, 24 and 48 hours. Those time points were selected to better gauge the specific genetic traits underpinning the sperm with improved fitness phenotype and the haploid selection process.

[0110] Across all samples, the loci deviating from the expected 50:50 ratios are predominantly present in autosomal protein coding genes and IncRNAs. Mechanisms of natural selection acting on protein coding genes are known (Bustamante et al., 2005), however, it was also found that approximately one third of the loci overlap with IncRNAs. This indicates that selection is likely to act not only on coding genes but also on regulatory elements such as IncRNAs. In fact, the findings presented here concord with previous findings in mice that reveal that IncRNAs exhibit similar diversity, divergence and signs of selection as protein coding genes (Wiberg et al. 2015). Therefore, protein coding genes and IncRNAs are likely to be direct targets of haploid selection and as such be suitable candidates for the development of genotypic markers.

[0111] Overall, a genome-wide concordance was observed across different assays, donors and selection time points. These observations demonstrate that natural selection is likely to favour alleles providing even the smallest phenotypic advantage to sperm cells. In fact, when selection acting on such molecular mechanisms acts alongside cellular events (e.g. sperm competition), slight differences in sperm performance are amplified in effect, driving genetic divergence between cells and affecting the offspring's fitness (as reported by Alavioon et al, 2017).

[0112] With regards to the genotypes under potential selection underpinning improved sperm fitness phenotypes, it has been shown that these are enriched in 50.6-81.7% of genes, which exhibit strong expression in the brain (Figure 1) involved in neuron projections, synapse and cell junction organisation (Figure 4). It is shown that sperm selected by the two selection assays present genetic signals coding for brain functions. In fact, it is shown that using the first (2 hour) selection assay resulted in more genetic hits for brain specific functions compared to the second selection assay selecting cells at 2 and 4 hours. Such differences can be expected as each human ejaculate is genotypically diverse. In addition, the differences can be attributed to the technical design differences in the selection assays. Using the second selection assay, fewer allelic differences between the sperm pools are found, yet these appear to have a greater genetic contribution to coding for brain specific function, particularly the samples selected after 12 hours of in vitro culture. Neuronal-like functions expressed were observed in pools selected in earlier time points (4, 8, 12 hours). However, these were also enriched in many other housekeeping functions such as ion channel complex and cell-cell adhesions. Nonetheless, these findings suggest that as time progresses, selection for alleles involved in brain function is particularly strong. However, sperm pools selected at 48 hours did not differ with regard to the enriched cellular functions compared to those selected at 24 hours. Enrichment for neuronal synapsis was prevalent in both sperm pools. Equally, both pools selected at 24 and 48 hours have approximately the same cell concentration, suggesting that phenotypes and genotypes are unlikely to change under selection once the optimum genotypes are achieved. Minimal gene overlap was observed between samples, donors and assays, indicating that there are likely to be multiple favourable allelic combinations that contribute to the same cellular and molecular functions. But five genes were identified in all donors, which are involved in tumor suppression as well as neurodevelopmental disorders, ASD and schizophrenia.

[0113] It is believed that this is the first investigation directly linking sperm improved fitness phenotype with brain-specific functions. Such observations suggest that the genes involved in these networks have pleiotropic effects. Shared functional genes between sperm and neurons are likely to have the same initial downstream targets but have subsequent different targets, giving rise to different cellular activities. Disruptions of such common pathways could lead to simultaneous impairment of both reproductive and neurodevelopmental processes (several lines of indirect evidence suggest that infertile couples using assisted reproductive technologies to conceive, sire offspring with a potential increased risk of developing neurodevelopmental conditions (Hart and Norman, 2013)). Both sperm and brainspecific tissues such as neurons execute high-energy demanding processes that require a great amount of metabolic support. The presence of neuronal receptors in sperm has been shown to reflect sperm motility, capacitation and acrosome reaction potential (Ramfrez-Reveco et al., 2017).

[0114] Current clinical options rely on the separation of motile from immotile sperm cells. These techniques select sperm based on membrane integrity, density and surface charge, whereby the sperm are phenotypically assessed through motility assays or differential gradients (Mcdowell et al., 2014). The present findings provide the foundation for developing genotypic selection options for sperm with improved fitness through means of, for example, flow cytometry or antibody technology.

[0115] In summary, the present findings present the first line of evidence that human sperm functional performance is linked with its underlying haploid genome. It is shown that haploid selection within an ejaculate changes allele frequencies but does not skew sperm sex ratios as a result.

[0116] Example 2: Phenotypic effects of sperm selection

[0117] The effects of within-ejaculate selection at 4I1 and 24b post seminal plasma removal in comparison to raw, washed and swim-up selected sperm were investigated, to assess the effects of selection on sperm phenotypes known to be linked with fertilisation success, namely swimming speed, morphology, chromatin and DNA integrity.

[0118] Materials and methods

[0119] 35 samples were included from 35 self-reported healthy donors (18-35 years old) through the University of East Anglia sperm donation programme (ethics no. 2020 / 2021-043).

[0120] The effects of selection on sperm longevity on multiple sperm traits were investigated, including sperm kinematics, morphology, chromatin and DNA integrity, and compared these with the same traits in the same samples from the raw ejaculate through several preparation steps. To this end, the baseline phenotypic variation in the raw ejaculated samples (R) was assessed, and then a swim-up was performed on a R subsample in a 15ml conical centrifuge tube (To). To control for potential interactions with the seminal plasma and avoid the formation of reactive oxygen species generated from potential leukocytes present in the seminal plasma, the remaining R sample was washed and the sperm traits assessed following washing (W). Afterwards, the W sample was split into two halves (T4 and T24) and incubated for 4 and 24 hours respectively. Finally in order to assess the effect of haploid selection, motile sperm were selected from T4 and T24 by a swim-up.

[0121] The swim-ups were performed on raw (n=25), T4 (n=35) and T24 (n=i2) samples as follows: 1 ml of human tubal fluid (HTF+) medium containing (in mM): 72.8 NaCl, 4.69 KC1, 0.2 MgSO4, 0.37 KH2PO4, 2.04 CaC12, 0.33 sodium pyruvate, 21.4 sodium lactate, 2.78 glucose, 21 HEPES, 25 NaHCOs and 3.5% human serum albumin (HSA) (Fisher, UK) adjusted to pH 7.4 with HCL was added to the semen / sperm pellet. During the swim-up assay, cells were allowed to swim up for th at 37°C with 5% CO2 in air. Up to 7OOpl of medium was then collected from the top of the assay to isolate the motile sperm.

[0122] Assessment of sperm concentration and swimming speed

[0123] Sperm swimming speed and concentration were assessed with the motility module of Integrated Semen Analysis System (ISAS) vi and the data was captured with an ISAS 782M camera connected to a UB2O3i microscope with a tox phase contrast objective and a pre-heated stage.

[0124] Assessment of sperm morphology

[0125] Morphology smears were prepared in duplicates using i5pl of sample smeared onto a slide and left to air dry at room temperature. The air-dried slides were stained with SpermBlue according to the manufacturer’s instructions. At least 200 cells were assessed under a UB2O3i microscope using brightfield microscopy at toox magnification and the percentage of broad overall normal sperm forms (as described in the 6th WHO manual (World Health Organization, 2021)).

[0126] Assessment of sperm chromatin and DNA integrity

[0127] Sperm chromatin condensation (AB) and DNA integrity (TB) was assessed using aniline and toluidine blue protocols. After mounting with Eukitt medium and coverslips, AB and TB slides were examined using a UB2O3i microscope with toox b rightfield objective. At least 200 sperm were assessed, and the percentage of immature chromatin (dark blue) and fragmented DNA (deep violet) were calculated (compromised cell number / total cell number *100%)

[0128] Statistical analyses

[0129] All analyses were performed in R V4.1.0 (R Core Team, 2021) and visualised using package ggplot2 V3.2.1. For all models involving sperm kinematics, a linear mixed effect model was performed to investigate the relationship between selection treatment and sperm pools using the Imer function from package lme4 vi.l.

[0130] Results

[0131] Mean values for baseline ejaculate and parameters (concentration, volume, pH, morphology and vitality) were above the lower reference values recommended by WHO (WHO, 6thedition, 2021). In general, all parameters displayed variation across the different donors.

[0132] Effect of selection on sperm motility (Figure 6A)

[0133] The sperm kinematics of Raw and Washed sperm pools were assessed and found no systematic evidence of improvements in the Washed pools in any of the kinematics parameters except wobble. Sperm collected from either of the Raw or Washed pools are expected to be highly heterogenous, as it is known that removing seminal plasma primarily aids in minimising immature germ cells, leukocytes and epithelial cells from the ejaculate, without selecting for a specific sperm trait. The results obtained confirm this trend. Next, the effects of swim-up were assessed and the sperm behaviour between Raw and To swim-up samples comparatively evaluated (Figure 6A). It was observed that curvilinear velocity (VCL), straight line velocity (VSL), average path velocity (VAP), straightness (STR), wobble (WOB), beat cross frequency (BCF) and lateral head displacement (ALH) but not linearity (LIN) were significantly higher in To swim-up sperm pools compared to Raw and Washed sperm pools.

[0134] Next, the effects of sperm selection were investigated and a clear increase in VCL, VSL, VAP, BCF and ALH was found, but not in LIN, STR, and WOB in the T4 samples selected by swim-up compared to Washed samples. Equally, an increase in VCL, VSL, VAP, STR, BCF, ALH was observed, but not in LIN or WOB in the T24 samples compared to Washed samples.

[0135] Following this, the question was asked whether selection following incubation improves sperm swimming speed compared to the swim-up without incubation (To) method traditionally used in clinical andrology laboratories. It was found that selection for sperm by swim-up following at 4I1 incubation increases most kinematic parameters except VSL, LIN, STR, WOB, BCF and ALH compared to To samples. Similarly, selection for sperm by swim-up following 24b incubation increases all kinematics, except LIN, STR, WOB and ALH compared to W samples.

[0136] Finally, systematic differences in sperm kinematics between T4 and T24 selected sperm pools were investigated, with no differences found between the two treatments in none of the direct or derived kinematic metrics.

[0137] Effect of selection on sperm morphology

[0138] Further assessments of the differences between sperm pools were performed by assessing gross sperm morphology (Figure 6C). Firstly, no difference was found in normal forms between the unselected sperm pools but a statistically significant difference between R and To sperm pools. Then, the W was compared to the selected sperm pools, which showed that selection at T4 and T24 had preferentially selected for sperm with better gross sperm morphology. Equally, T4 selected samples had greater number of morphologically normal cells compared to To samples and T24. Finally, no significant differences were found in morphology between T4 and T24 samples.

[0139] Effect of selection on chromatin and DNA integrity

[0140] When comparing chromatin compaction among sperm pools, it was found that the levels of chromatin maturity were similar between the Raw and Washed sperm pools, but not between R and To samples (Figure 6B). When W was compared to T4 and T24 samples, a significantly higher number of cells with mature chromatin was found in both T4 and T24 sperm pools compared to Washed pools (Figure 6B). Similarly, a higher number of cells with mature chromatin in T4 and T24 pools were found compared to To samples. Within the selection for longevity treatment (i.e. T4 vs T24), chromatin maturity did not differ. Similarly, when comparing DNA integrity among R and W sperm pools, no significant differences were found (Figure 6D). However, the To samples had greater DNA integrity compared to R samples. In contrast to W sperm pools, significantly higher DNA integrity was observed in the pools that were exposed to the T4 selection treatment, but not to T24 treatment. Interestingly, the To samples had greater DNA integrity compared to T4 and T24 samples. Lastly, similar levels of DNA integrity were found in the longevity treatment groups (Figure 6D).

[0141] Discussion

[0142] This Example suggests that sperm selection following incubation in humans eliminates most sperm with phenotypic defects from an ejaculate. Key sperm traits associated with fertilisation success were assessed in raw, washed and swim-up samples and compared to traits observed in sperm selected at four and 24 hours following incubation by swim-up. Striking improvements in velocity, morphology, chromatin and DNA integrity were found in T4 and T24 samples compared to sperm in unselected sperm pools. However, the To samples had greater DNA integrity compared to T4 and T24 samples. This study provides evidence that the selection assay described in this study (incubation for 4 (T4) or 24 (T24) hours followed by swim-up) outperforms swim-up without incubation (To) across most sperm traits, including sperm motility, morphology and chromatin maturity. This study also demonstrates that DNA integrity decreases as a result of the incubation period (T4, T24) compared to To.

[0143] Sperm selection using current methods can be practically challenging in clinical or agricultural settings. In addition, during ART it can take up to five hours from gamete collection to fertilisation. In these scenarios, sperm preparations are often kept at room temperature, compromising the sperm fertilisation ability. In addition, during ART many, if not all, selective barriers are by-passed, raising potential questions about the nature of sperm used for fertilisation, especially when within ejaculate genetic and epigenetic heterogeneity have been previously reported in human samples. The results presented here indicate that incubating sperm prior to selection (thereby selecting the ‘fittest’ sperm) significantly improves many of the traits associated with fertilisation success and embryo development. The results here also indicate that subjecting sperm to an incubation period negatively affects DNA integrity. These results demonstrate that there would be significant advantages to being able to identify sperm for use in an ART setting based on markers expressed by the fittest sperm, as this would allow the fittest sperm to be selected from a sample, whilst avoiding the need to subject the sample to an incubation period.

[0144] Example a: Human Sperm Proteomics

[0145] Four human sperm ejaculates were collected: two from the University of East Anglia (UEA), Norwich, UK; two from Cape Town, South Africa (ZA)).

[0146] Samples were analysed for volume and overall sperm motility:

[0147] Samples were washed in human tubal fluid (HTF) and condensed to 200 pL. Sperm were then selected from the samples using two different assays, as detailed below.

[0148] In vitro sperm selection assays

[0149] (i) The two samples collected from the University of East Anglia underwent a methylcellulose sperm selection assay (similar to the first sperm selection assay set out in Example 1): The condensed 200 pL samples from the first and second UEA donors were placed in the centre of separate concave watch glasses filled with 10 mL of human tubal fluid (HTF) with 3.5 % human serum albumin and 1% methyl cellulose. The samples sank and filled a circle at the bottom of the centre of the glass in each case. The watch glasses were then incubated at 37°C with 5% C02at a constant humidity for 4 hours. After incubation, 4 mL of sample was removed from outside of a 2cm diameter within edge watch glass (edge sample), followed by the removal of 2 mL of sample from the centre of each watch glass (middle sample). The middle sample was collected 0.5 cm above the bottom of the watchglass in order to collect sperm that were alive and motile at the beginning of the assay and which were capable of penetrating the methyl cellulose by this distance, thereby avoiding contamination with dead sperm and somatic cells. The samples were washed with 1 ml of human tubal fluid per sample. The samples were then centrifuged for 10 mins at 800 rpm and the supernatant removed. Each pellet was resuspended in HTF and the samples analysed for cell number and motility:

[0150] Each sample was then condensed to i mL and treated with cellulase at 50 °C for 1 hour. Each sample was then pelleted by centrifugation at 10,000 rpm for 5 minutes, the supernatant was removed and samples were flash frozen using liquid nitrogen.

[0151] (ii) The two samples collected from South Africa underwent a swim-up sperm selection assay (similar to the second sperm selection assay set out in Example 1): The fresh, raw ejaculate from each donor was divided into two samples. One sample from each donor underwent a washing step followed by a swim up assay as follows: each sample was aliquoted into 4 Eppendorf tubes (5Opl in each tube; this was done in order to improve the efficiency of the assay and collection of sufficient sperm for mass spectrometry, as using 4 tubes helped avoid motile sperm blocking each other at the top of the tube), topped with iml of HTF. The samples were allowed to swim up for 45 minutes and 5OOpl from the upper portion of each tube was aliquoted and transferred to a new Eppendorf tube. The 4 aliquots for each donor were then mixed, centrifuged for 10 mins at 800 rpm and the supernatant removed. Each pellet was resuspended in imL HTF and the samples analysed for cell number and motility (To sample).

[0152] The other sample from each donor was washed and re-suspended in 200pl of human tubal fluid with 3.5 % human serum albumin and incubated at 37°C with 5% C02at a constant humidity for 4 hours and then underwent a swim up assay as detailed above (T4 sample).

[0153] Protein Extraction & Proteomics Protein was extracted from all collected samples: samples were thawed and 450 pl of acetone was added to each of the thawed samples. Samples were then vortexed for 15 minutes, then centrifuged at 10,000 rpm for 10 minutes. Supernatant was then removed and 500 pl of acetone added to each sample. Each sample was then spun gently for two minutes manually then centrifuged for 15 mins at 10000 rpm. Supernatant was removed and samples were air dried for 1 and a half hours.

[0154] The protein extracts from each sample were analysed for proteomics using separate Orbitrap TMT quantitative mass spectrometry. The sample from South Africa donor 1 (swim-up assay) was re-analysed on a separate TMT panel with 7 replicate runs due to the low significance found in the first panel.

[0155] Further analysis was then undertaken: Protein pellets were resuspended in 100 pl of 2.5% sodium deoxycholate in 0.2 M EPPS-buffer, pH 8.5, and the samples vortexed under heating. The samples were treated with dithiothreitol and iodoacetamide to alkylate cysteine residues and digested with trypsin in the SDC buffer according to standard procedures. The SDC was then precipitated by adjusting to 0.2% TFA, and the clear supernatant subjected to C18 SPE (Reprosil-Pur 120 C18-AQ, 5 um). Peptide concentration was estimated by running an aliquot of the digests on LCMS (see below). TMT labelling was performed using 8 channels from a TMT™ioplex kit (TMT lot: VH306773, ThermoFisher Scientific) according to the manufacturer’s instructions with slight modifications; samples were dissolved in 90 pl of 0.2 M EPPS buffer (MERCK) / io% acetonitrile, and 200 pg TMT reagent dissolved in 22 pl of acetonitrile was added. Samples were assigned to the TMT channels in two separate experiments, according to Table X, below. After 2 h incubation, aliquots of 1.5 pl from each sample from each individual TMT experiment were combined in 250 pl 0.2% TFA, desalted, and analysed on the mass spectrometer to check the labelling efficiency and estimate total sample abundances. The main sample aliquots were quenched by adding 8 pl of 5% hydroxylamine and combined to roughly level abundances. The peptides were desalted using a C18 Sep-Pak cartridge (200 mg). The eluted peptides were dissolved in 500 pl of 25 mM NH4HCO3and fractionated by high pH reversed phase HPLC. The samples were loaded to an XBridge® 3.5 pm C18 column (150 x 3.0 mm, Waters). Fractionation was performed on an ACQUITY Arc Bio System (Waters) with the following gradient of solvents A (water), B (acetonitrile), and C (25 mM NH4HCO3in water) at a flow rate of 0.5 ml min1: solvent C was kept at 10% throughout the gradient; solvent B: 0-5 min: 5%, 5-10 min: 5-10%, 10-60 min: 10-40%, 60-75 min: 40-80%, followed by 5 min at 80% B and re-equilibration to 5% for 24 min. Fractions were collected every 1 min and concatenated by combining fractions of similar peptide concentration to produce 20 final fractions for MS analysis. A replicate of one donor sample (2 TMT channels) was fractionated using the Pierce™ High pH Reversed-Phase Peptide Fractionation Kit (Thermo) to produce 7 fractions.

[0156] Aliquots of all fractions were analysed by nanoLC-MS / MS on an Orbitrap Eclipse™ Tribrid™ mass spectrometer equipped with a FAIMS Pro Duo interphase coupled to an UltiMate® 3000 RSLCnano LC system. The samples were loaded onto a trap cartridge (PepMap™ Neo 5 pm C18300 pm X 5 mm Trap Cartridge, Thermo) with 0.1% TFA at 15 pl min1for 3 min. The trap column was then switched in-line with the analytical column (Aurora Frontier TS, 60 cm nanoflow UHPLC column, ID 75 pm, reversed phase C18, 1.7 pm, 120 A; lonOpticks, Fitzroy, Australia) for separation using the following gradient of solvents A (water, 0.1% formic acid) and B (80% acetonitrile, 0.1% formic acid) at a flow rate of 0.23 pl min1: 0-3 min 1% B (parallel to trapping); 3-10 min linear increase B to 9 %; 10-105 min increase B to 50%; followed by a ramp to 99% B and re-equilibration to 1% B.

[0157] Data were acquired with the following parameters in positive ion mode: MS1 / 0T: resolution 120K, profile mode, mass range m / z 400-1600, AGC target 4e5, max inject time 50 ms, FAIMS device set to three compensation voltages (-35V, -50V, - 65V) for 1 s each; MS2 / IT: for each CV, data dependent analysis with the following parameters: 1 s cycle time Rapid mode, centroid mode, quadrupole isolation window 0.7 Da, charge states 2-5, threshold i-9e4, CID CE=33, AGC target ie4, max. inject time 50 ms, dynamic exclusion 1 count for 15 s with mass tolerance of 7 ppm; MS3 synchronous precursor selection (SPS): 10 SPS precursors, MS2 isolation window 0.7 Da, HCD fragmentation with CE=6s, Orbitrap Turbo TMT and TMTpro resolution 30k, AGC target 200%, max inject time 100 ms; Real Time Search (RTS): protein database UPooooo564O_human_2O24_96o6.fasta (Uniprot, 20,609 entries)., enzyme trypsin, 1 missed cleavage, oxidation (M) as variable, carbamidomethyl (C) and TMT6plex as fixed modifications, precursor tolerance 10 ppm, Xcorr = 1.4, dCn = 0.1.

[0158] The acquired raw data were processed and quantified in Proteome Discoverer 3.1 (Thermo Fisher Scientific); all mentioned tools of the following workflow are nodes of the proprietary Proteome Discoverer (PD) software.

[0159] The UPooooo564O_human_2O24_96o6.fasta protein fasta database (Uniprot, 20,609 entries) was imported into PD adding a reversed sequence database for decoy searches. The database search was performed using the incorporated search engines CHIMERYS (MSAID, Munich, Germany) and Comet.

[0160] An unselected sample (middle sample) and a selected sample (edge sample) from each of the UEA donors underwent quantification and protein ratio calculation. In addition, an unselected (To) sample and a selected (T4) sample collected from each of the ZA donors underwent protein ratio calculation.

[0161] The processing workflow for both search engines included recalibration of MS1 spectra (RC), reporter ion quantification by most confident centroid (20 ppm) and a search on the human protein database (as imported above). For CHIMERYS the Top N Peak Filter was applied with 20 peaks per too Da. Then the infeiys _3.o.o._fragmentation prediction model was used with fragment tolerance of 0.3 Da, enzyme trypsin with 1 missed cleavage, variable modification oxidation (M), fixed modifications carbamidomethyl (C) and TMT6plex on N-terminus and K. For Comet the version 2019.01 rev.o parameter file was used with default settings except precursor tolerance set to 6 ppm and trypsin missed cleavages set to 1. Modifications were the same as for CHIMERYS.

[0162] The consensus workflow included the following parameters: intensity-based abundance, normalisation on total peptide abundances, protein abundance-based ratio calculation, only unique peptides (protein groups) for quantification, TMT channel correction values applied (Lot VH306773), co-isolation / SPS matches / CHIMERYS Coefficient thresholds 50% / 70%, 0.8, missing values imputation by low abundance resampling, hypothesis testing by t-test (background based), adjusted p-value calculation by BH-method. The results were exported into a Microsoft Excel table including data for normalised and un-normalised abundances, ratios for the specified conditions, the corresponding p-values and adjusted p- values, number of unique peptides, q-values, PEP- values, identification scores from both search engines; FDR confidence filtered for high confidence (strict FDR 0.01) only. Results

[0163] The initial TMT quantitative proteome produced for the sperm cells that underwent the methyl cellulose assay (UEA donors 1 and 2) found 7528 protein matches. These results were filtered to remove single peptide matches and potential contaminants which produced a list of 6155 proteins, which could be confidently identified within the sample. Further analysis of these proteins found that a total of 255 displayed a log2-fold change of above 2 or below -2, with a significant p-value showing that these 255 proteins had significantly different abundances between the middle and edge samples.

[0164] TMT quantitative proteome analysis of sperm cells that underwent the swim-up assay (ZA donors 1 and 2) produced a protein list of 7256. This was filtered in a similar way to the proteomic data from the sperm cells that underwent the methyl cellulose assay to produce a list of 5898 proteins. Of this total list, 111 proteins were identified as having a log2-fold change of above 2 or below -2, with a significant p- value. However, these were not all significant in both samples as only 18 proteins in the proteome from the sample of ZA donor 1 showed a significant p-value after normalisation.

[0165] A repeat of the TMT analysis of swim-up assay samples from ZA donor 1 produced a total proteome of 3026 which was filtered to a final 2284 proteins. From this list, 30 proteins were identified as having a log2-fold change of above 2 or below -2, with a significant p-value. This data maybe considered as more robust for ZA donor 1 as the data was analysed across seven fractions instead of one, however as this was a separate method to the initial TMT experiment performed on the swim-up samples, they cannot be directly compared.

[0166] Whilst direct comparison of the TMT data from each experiment is not possible due to the differences in labelling, it is possible to compare which proteins were present and found to be significantly different in each sample as well as the trend in abundances between the control samples (Time-zero (To) or middle samples) and samples that underwent selection (Time- hours (T4) or edge samples) (Figure 7A- E). One protein was found to have a significant difference between the control and selected groups for all assays and all donors, with the same trend, which was Olfactomedin-4 (Uniprot ID: Q6UX06, OLFM4).

[0167] OLFM4 has previously been recognised to play a role in the cell cycle of various cancer cells, such as promoting proliferation in pancreatic cancer (Kobashi et al. 2007) and potentially inducting cell differentiation in myeloid leukemic cells (Liu et al. 2010), however, its exact role in human sperm is unknown. Without being bound by a particular theory, it is possible that as OLFM4 is involved in cell adhesion, it may cause sperm cells to ‘clump’ together and prevent progressive movement, therefore being detrimental at high abundances for sperm fitness.

[0168] Following initial analysis of all of the TMT data, a list of 77 potential sperm cellsurface biomarkers for sperm fitness was generated (Table A, below). These 77 proteins were found to be significantly different between control and selected samples for at least 2 of the donor samples, and present in at least 4 of the proteomes. These biomarkers were established using AlphaFold predictions for the presence of signal peptides and transmembrane regions as well as previously established cellular location data from other cell types using the Uniprot database. These biomarkers have been ranked according to presence and significance in the TMT proteomics data as well as the sperm pool DNA data.

[0169] A further, more stringent analysis of all the TMT data was then undertaken, with criteria requiring a consistent twofold change in the same direction across at least three donors, and a padj z < 0.05 in at least one donor. This analysis identified 634 proteins with significantly lower abundance and 544 proteins with significantly higher abundance in selected sperm. The combined set of 1,178 proteins showing differential abundance was further refined to generate shortlists of candidate biomarkers. For proteins with lower abundance in selected sperm, more stringent filtering was applied due to the strength of observed directional change. This resulted in a shortlist of 62 proteins. In contrast, although 544 proteins were initially identified with significantly higher abundance, these changes were less consistent across donors. Therefore, a looser criterion was applied, including only proteins with a fold-change ratio above 2 in at least two donors, and significance in at least one donor. This added a further six proteins to the shortlist. From the combined shortlist of 68 proteins, top candidates were selected using additional criteria.

[0170] The top six proteins with lower abundance in selected sperm (i.e. negative biomarkers, where low levels of expression indicates sperm with improved fitness) - OLFM4, CTNNG, PLA2, beta-defensin 105, hnRNP K and MK - were present in all five samples (four donors with one repeated), statistically significant in at least four donors, and had a median fold-change ratio below 0.2. Similarly, the top four proteins with higher abundance in selected sperm (i.e. positive biomarkers, where higher levels of expression indicates sperm with improved fitness) - thyroxine- binding globulin (TBG), Afamin, Beta-microseminoprotein (MSP) and Alpha-i-acid glycoprotein 1 (A1AG1; also known as orosomucoid (ORM)) - were present in all five donors, significant in at least four donors, and had a median fold-change ratio above 2.

[0171] Example 4: Biomarker validation

[0172] Five human sperm ejaculates were collected from donors from the University of East Anglia (UEA), Norwich, UK. Samples were analysed for volume and overall sperm motility:

[0173] Samples were then aliquoted to approximately 1.5 million cells in 1 mL of HTF medium. Aliquots were incubated for 4 hours at 37 °C with 5% C02.

[0174] Following incubation, samples were re-assessed for motility and cell number, then immediately stained with eFluor™ 520 fixable viability dye (eBioscience, Invitrogen) according to the manufacturer’s protocol. After viability staining, cells were fixed in 4% paraformaldehyde at 4 °C. Fixed cells were then blocked using Human Fc Receptor Blocking Solution to minimise non-specific antibody binding.

[0175] Fluorophore-conjugated primary antibodies were subsequently added and incubated under standard conditions. The following antibodies were used: anti-OLFMq clone 12, conjugated to R-phycoerythrin (PE), at a 1:10 dilution (Novus Biologicals, NBP3- 06415PE); and anti-CTNNG (CTNNG / 2155R), conjugated to mFluor Violet 450 SE, at 1:10 dilution (Novus Biologicals, NBP3-08485MFV450). Throughout staining, cells underwent minimal washing steps to preserve cell number and were stored overnight at 4 °C prior to flow cytometry analysis.

[0176] Samples were analysed using the BD FACSDiscover™ S8 Cell Sorter, operated in fully spectral mode. No cell sorting was performed. Imaging and fluorescence data were acquired simultaneously for each sample with fixable viability dye (eFluor™ 520) detected using imaging laser Bi. Events were gated to isolate the sperm population (i.e. to exclude debris and aggregates) (Pi), based on forward and side scatter properties .All subsequent data acquisition and analysis were restricted to this Pi population. For each antibody-stained sample, between 5,000 and 50,000 individual events within the Pi gate were recorded, depending on sample quality and cell availability.

[0177] Spectral unmixing was performed using BD FACSDiva™ software based on singlestained controls for each fluorophore-conjugated antibody and the fixable viability dye. An unstained sperm sample was included to control for cellular autofluorescence during unmixing. Compensation and unmixing matrices were applied prior to further analysis.

[0178] Flow cytometry image files were merged with the primary fluorescence data using BD Cellview™ Image Extractor. The resulting datasets were processed in R V4.4.3. Data were normalised using the CytoNorm package, and quality control was performed using PeacoQC to identify and remove anomalous events. Gated and cleaned data were subsequently analysed in FlowJo v 10.10.0 (BD Life Sciences), with all events restricted to the Pi sperm population. Statistical modelling of marker expression in live and dead cell populations was conducted in R V4.4.3, with residual diagnostics simulated using the DHARMa package to evaluate model fit and distribution assumptions.

[0179] Results

[0180] Following data acquisition, gating was initially applied using one representative sample (GD219) to separate live and dead cells. The eFluor 520 signal enabled clear resolution between these populations. For each of OLFM4 and CTNNG, a distinct band of cells located between the live and dead gates was observed; these likely represent apoptotic cells at the time of fixation, exhibiting intermediate staining intensity. A fluorescence intensity threshold level of around io5arbitrary units was used for OLFM4 and CTNNG. This threshold level reflected the minimum signal at which sperm expressing the selected biomarkers could be reproducibly separated from the rest of the population.

[0181] OLFM4

[0182] A heatmap colour scale was applied to the OLFM4 signal, and a fluorescence intensity threshold of to5was selected to distinguish OLFM4-positive cells. Cells with a fluorescence intensity above this value were classified as OLFM4 positive. The proportions of cells in each category (OLFM4+ / - live / dead) are summarised in Table C, below.

[0183] Across all five samples, the proportion of OLFM4-positive cells was significantly higher in the dead population compared to the live population. This was tested using a binomial generalised linear mixed model with the proportion of OLFM4-positive cells as the response variable, live / dead status as a fixed effect, and sample ID as a random intercept. Model diagnostics performed using the DHARMa package indicated no evidence of overdispersion, zero-inflation, or model misspecification.

[0184] CTNNG

[0185] A fluorescence intensity threshold of to5was selected for CTNNG positivity, based on visual inspection of the heatmap colour scale. Cells with intensities above this value were classified as CTNNG-positive. The proportions of CTNNG+ / - cells within live and dead populations are shown in Table C. This threshold was not formally validated but allowed for clear visual discrimination across the dataset.

[0186] Across all five samples, the proportion of CTNNG-positive cells was higher in the dead population compared to the live population, as assessed by a binomial generalised linear mixed model with live / dead status as a fixed effect and sample ID as a random intercept. The effect was significant, with a p-value ofp < 0.001. DHARMa diagnostic plots indicated a well-fitted model with no violations of model assumptions.

[0187] Table C

[0188] These findings support the proteomic results of Example 3, with OLFM4 and CTNNG demonstrated to be significantly more abundant in less fit sperm.

[0189] Table A

[0190]

[0191]

[0192]

[0193]

[0194]

[0195]

[0196]

[0197] Table B

[0198]

[0199]

[0200]

[0201] In addition, there are provided the following technological aspects:

[0202] 1. A method of identifying sperm comprising one or more alleles associated with improved fitness within a sample of sperm, the method comprising: providing a sperm sample; and a) detecting sperm in the sample which express one or more biomarkers associated with the one or more alleles; or b) detecting sperm in the sample which express at least a log2two-fold change in expression level of one or more biomarkers associated with the one or more alleles relative to the other sperm within the sample; or c) detecting sperm in the sample which express at least a threshold level of one or more biomarkers associated with the one or more alleles; or d) detecting sperm in the sample which do not express one or more biomarkers associated with the one or more alleles; wherein the one or more biomarkers include OLFM4. The term ‘other sperm’ as used in part b) refers to sperm within the sample which express the lowest or highest level (as appropriate for that biomarker) of the biomarker in question.

[0203] 2. A method according to clause 1, wherein the one or more biomarkers further comprise one or more of phospholipase A2 (PLA2), beta-defensin 105 (DEFB105A), heterogeneous nuclear ribonucleoprotein K (hnRNP K or protein K) and Midkine (MK).

[0204] 3. A method according to clause 1 or 2, wherein the change in expression level of the one or more biomarkers relative to the other sperm in part b) is: i. a decrease in expression; or ii. an increase in expression.

[0205] 4. A method according to any one of clauses 1 to 3, wherein the change in expression level of the one or more biomarkers relative to the other sperm in part b) is at least log2two-fold, or at least log2three-fold, at least log2four-fold, at least log2five-fold.

[0206] 5. A method according to any one of clauses 1 to 4, wherein the sperm sample is from a human or non-human mammal.

[0207] 6. A method according to any one of clauses 1 to 5, wherein the sperm sample is from a human.

[0208] 7. A method according to any one of clauses 1 to 6, wherein the sperm sample is fresh; or is a frozen sample that has been thawed.

[0209] 8. A method according to any one of clauses 1 to 7, wherein the method does not comprise selecting sperm on the basis of carrying an X or Y chromosome.

[0210] 9. A method of selecting sperm identified according to any one of clauses 1 to 8, comprising separating a target population from the sperm sample, wherein: a) the sperm in the target population substantially comprise or consist of sperm which express one or more of the biomarkers, and the target population constitutes the selected sperm; or b) the sperm in the target population substantially comprise or consist of sperm which have at least a log2two-fold change in expression level of one or more of the biomarkers relative to the other sperm within the sample; and the target population constitutes the selected sperm; or c) the sperm in the target population substantially comprise or consist of sperm expressing at least a threshold level of one or more of the biomarkers; and the target population constitutes the selected sperm; or d) the sperm in the target population substantially comprise or consist of sperm which do not express one or more of the biomarkers; and the target population constitutes the selected sperm; or e) the sperm in the target population substantially comprise or consist of sperm which do not express a threshold level of one or more of the biomarkers; and the target population constitutes the selected sperm; or f) the sperm in the target population substantially comprise or consist of sperm which do not express one or more of the biomarkers, or do not have at least a log2two-fold change in expression level of one or more of the biomarkers relative to the other sperm within the sample, or do not express a threshold level of one or more of the biomarkers; and the sperm remaining after separation of the target population constitute the selected sperm.

[0211] 10. An enriched population of sperm comprising the one or more alleles associated with improved fitness, selected according to the method of clause 9.

[0212] 11. A method for assessing the reproductive potential of a sample of sperm, comprising: i) quantifying the number of sperm in a target population selected using the method according to clause 9; or ii) assessing the ratio of sperm in a sample which are in a target population selected using a method according to clause 9, to the sperm within the sample which are not in the target population.

[0213] 12. A method for identifying sperm with one or more alleles associated with improved fitness within a sample of sperm, comprising comparing the expression of one or more biomarkers associated with the one or more alleles to a pre-determined standard for each of the one or more biomarkers, wherein a significant difference in expression of the one or more biomarkers in the sample as compared to the pre-determined standard for each of the one or more biomarkers indicates that the sperm has one or more alleles associated with improved fitness; wherein the one or more biomarkers include OLFM4.

[0214] 13. A method according to clause 12, wherein the one or more biomarkers further comprise one or more of phospholipase A2 (PLA2), beta-defensin 105 (DEFB105A), heterogeneous nuclear ribonucleoprotein K (hnRNP K or protein K) or Midkine (MK).

[0215] 14. The method of any one of clauses 1 to 9 or clause 11 to 13, or the enriched population of sperm according to clause 10, for use in an assisted reproductive technology setting.

[0216] The various embodiments described herein are presented only to assist in understanding and teaching the claimed features. These embodiments are provided as a representative sample of embodiments only and are not exhaustive and / or exclusive. It is to be understood that advantages, embodiments, examples, functions, features, structures, and / or other aspects described herein are not to be considered limitations on the scope of the invention as defined by the claims or limitations on equivalents to the claims, and that other embodiments may be utilised, and modifications may be made without departing from the scope of the claimed invention. Various embodiments of the invention may suitably comprise, consist of, or consist essentially of, appropriate combinations of the disclosed elements, components, features, parts, steps, means, etc, other than those specifically described herein. In addition, this disclosure may include other inventions not presently claimed, but which may be claimed in future.

Claims

Claims1. A method for identifying a population of sperm with one or more alleles associated with improved fitness within a sample of sperm, comprising comparing the expression of one or more biomarkers associated with the one or more alleles to a threshold level for each of the one or more biomarkers; wherein the one or more biomarkers include Olfactomedin-4 (OLFM4) and / or CTNNG.

2. A method as claimed in claim 1, wherein at least one of the biomarkers is a negative biomarker, and sperm with improved fitness express below the threshold level of the biomarker.

3. A method as claimed in claim 2, wherein the population of sperm comprising one or more alleles associated with improved fitness in the sample express below the threshold level of OLFM4 and / or CTNNG.

4. A method as claimed in claim 1, wherein at least one of the biomarkers is a positive biomarker, and sperm with improved fitness express at least the threshold level of the biomarker.

5. A method as claimed in any one of claims 1-4, wherein the one or more biomarkers further comprise one or more of phospholipase A2 (PLA2), beta-defensin 105 (DEFB105A), heterogeneous nuclear ribonucleoprotein K (hnRNP K or protein K), Midkine (MK), thyroxine-binding globulin (TBG), Afamin, Beta-microseminoprotein (MSP) and Alpha-1- acid glycoprotein 1 (A1AG1).

6. A method as claimed in claim 5, wherein the population of sperm comprising one or more alleles associated with improved fitness in the sample express below the threshold level of PLA2, DEFB105A, hnRNP K and / or MK; and / or at least the threshold level of TBG, Afamin, MSP and / or A1AG1.

7. A method as claimed in any one of claims 1-6, wherein the threshold reflects an expression level of at least a log2two-fold change relative to a control.

8. A method as claimed in any one of claim 7, wherein the threshold reflects anexpression level of at least a log2three-fold, at least log2four-fold, at least log2five-fold, or at least log2six-fold change relative to a control.

9. The method as claimed in any one of claims 1-8, further comprising determining the sperm population expressing below the threshold level of at least one of the biomarkers to be a first population; and the sperm population expressing at least the threshold level of the same biomarker(s) to be a second population.

10. A method as claimed in any one of claims 1 to 9, wherein the sperm sample is from a human or non-human mammal, optionally wherein the sperm sample is from a human.

11. A method as claimed in any one of claims 1 to 10, wherein the method does not comprise selecting sperm on the basis of carrying an X or Y chromosome.

12. A method of selecting a population of sperm identified according to any one of claims 9 to 11, comprising:(a) separating the first population from the sperm sample, wherein the sperm in the first population constitutes the selected sperm; or(b) separating the second population from the sperm sample, wherein the second population constitutes the selected sperm.

13. A method for assessing the reproductive potential of a sample of sperm, comprising: i) quantifying the number of sperm in a population selected using the method according to claim 12; or ii) assessing the ratio of sperm in a population selected using a method according to claim 12, to the sperm within the sample which are not in the selected population.

14. The method of any one of claims 1 to 13 for use in an assisted reproductive technology setting.

15. An enriched population of sperm comprising the one or more alleles associated with improved fitness, selected according to the method of claim 12, optionally wherein the enriched population of sperm are for use in an assisted reproductive technology setting.

Images