METHOD FOR DETERMINING THE QUALITY OF AN ANIMAL'S SEMEN

Abstract

A method is for determining the quality of an animal's semen. The method includes the steps of collecting at least one fresh or frozen semen sample, and measuring at least one absorption spectrum X.sub.j of at least one sample of the semen. The sample is collected to a straw for artificial insemination with animal semen obtained by implementing the method. A computer and software are used for the implementation of the method.

Claims

1. A method of determining the quality of an animal's semen, comprising the steps of: collecting at least one fresh or frozen semen sample, measuring at least one absorption spectrum of at least one sample of said semen, determining from said at least one absorption spectrum X.sub.j, a value of the first derivative of the absorptions X′.sub.j, calculating one or more parameters, representative of the quality of said semen, selected from the group consisting of concentration Y1, mobility Y2, progressive spermatozoa rate Y3, viability Y4, percentage of live spermatozoa with stable phospholipids Y5, mitochondrial potential Y6, percentage of spermatozoa with peroxidised lipids Y7, percentage of spermatozoa with an intact acrosome Y8, the total antioxidant capacity TAC Y9, fatty acid composition Y10, percentage of spermatozoa having a normal morphology Y11, the osmolarity Y12 and glutathione GSH level Y13, non-return rate at 56 days Y14, the non-return rate at 90 days Y15, pregnancy diagnosis Y16, from said first derivative of the absorption X′.sub.j; previously determined for the determining of the quality of said semen.

2. The method according to claim 1, wherein said calculation is a calculation of at least two of said parameters.

3. The method according to claim 1, wherein said at least one absorption spectrum X.sub.j comprises a first wavenumber range selected in the wavenumber range [1,800 cm−1; 900 cm−1] and/or a second wavenumber range selected from the wavenumber range [3,000 cm.sup.−1; 2,700 cm.sup.−1].

4. The method according to claim 1, wherein: said concentration Y1 is calculated according to the mathematical law Y1=β.sub.0+Σ.sub.j=1.sup.nβ.sub.jX′.sub.j, where X′.sub.j(j∈[1;n]) is the first derivative of the absorption X.sub.j, normalised by an SNV (“Standard Normal Variate”) pre-processing, for the wavenumber range [1,800 cm.sup.−1; 900 cm.sup.−1] and the weighting coefficients β.sub.0 and β.sub.j(j∈[1;n]) are constants; and/or said mobility Y2 is calculated according to the mathematical law Y2=β.sub.0+Σ.sub.j=1.sup.nβ.sub.jX′.sub.j where X′.sub.j(j∈[1;n]) is the first derivative of the absorption X.sub.j, preferably normalised by an SNV pre-processing, for the wavenumber range [1,800 cm−1; 900 cm−1], and the weighting coefficients β.sub.0 and β.sub.j(j∈[1;n]) are constants; and/or said progressive spermatozoa rate Y3 is calculated according to the mathematical law Y3=β.sub.0+Σ.sub.j=1.sup.nβ.sub.jX′.sub.j, where X′.sub.j(j∈[1;n]) is the first derivative of absorption X.sub.j, normalised by an SNV pre-processing, for the wavenumber range [1,800 cm−1; 900 cm−1], and the weighting coefficients β.sub.0 and β.sub.j(j∈[1;n]) are constants; and/or said viability Y4 is calculated according to the mathematical law Y4=β.sub.0+Σ.sub.j=1.sup.nβ.sub.jX′.sub.j, where X′.sub.j(j∈[1;n]) is the first derivative of the absorption X.sub.j, normalised by an SNV pre-processing, for the wavenumber range [1,800 cm−1; 900 cm−1] and the weighting coefficients β.sub.0 and β.sub.j(j∈[1;n]) are constants; and/or said percentage of live spermatozoa with stable phospholipids Y5 is calculated according to the mathematical law Y5=β.sub.0+Σ.sub.j=1.sup.nβ.sub.jX′.sub.j, where X′.sub.j(j∈[1;n]) is the first derivative of the absorption normalised by an SNV pre-processing, for the wavenumber range [1,800 cm−1; 900 cm−1] and the weighting coefficients β.sub.0 and β.sub.j(j∈[1;n]) are constants; and/or said mitochondrial potential Y6 is calculated according to the mathematical law Y6=β.sub.0+Σ.sub.j=1.sup.nβ.sub.jX′.sub.j, where X′.sub.j(j∈[1;n]) is the first derivative of the absorption X.sub.j, normalised by an SNV pre-processing, for the wavenumber range [1,800 cm−1; 900 cm−1] and the weighting coefficients β.sub.0 and β.sub.j(j∈[1;n]) are constants; and/or said percentage of spermatozoa with peroxidised lipids Y7 is calculated according to the mathematical law Y7=β.sub.0+Σ.sub.j=1.sup.nβ.sub.jX′.sub.j, where X′.sub.j(j∈[1;n]) is the first derivative of the absorption X.sub.j, normalised by an SNV pre-processing, for the wavenumber range [1,800 cm−1; 900 cm−1] and the weighting coefficients β.sub.0 and β.sub.j(j∈[1;n]) are constants; and/or said percentage of spermatozoa with an intact acrosome Y8 is calculated according to the mathematical law Y8=β.sub.0+Σ.sub.j=1.sup.nβ.sub.jX′.sub.j, where X′.sub.j(j∈[1;n]) is the first derivative of the absorption X.sub.j, normalised by an SNV pre-processing, for the wavenumber range [1,800 cm−1; 900 cm−1] and the weighting coefficients β.sub.0 and β.sub.j(j∈[1;n]) are constants; and/or said total antioxidant capacity TAC Y9 is calculated according to the mathematical law Y9=β.sub.0+Σ.sub.j=1.sup.nβ.sub.jX′.sub.j, where X′.sub.j(j∈[1;n]) is the first derivative of the absorption X.sub.j, normalised by an SNV pre-processing, for the wavenumber range [1,700 cm−1; 910 cm−1] and the weighting coefficients β.sub.0 and β.sub.j(j∈[1;n]) are constants; and/or said fatty acid composition Y10 is calculated according to the mathematical law Y10=β.sub.0+Σ.sub.j=1.sup.nβ.sub.jX′.sub.j, where X′.sub.j(j∈[1;n]) is the first derivative of the absorption X.sub.j, normalised by an SNV pre-processing, for the wavenumber range [1,800 cm−1; 900 cm−1] and for the wavenumber range [3,000 cm−1; 2,700 cm−1] and the weighting coefficients β.sub.0 and β.sub.j(j∈[1;n]) are constants; and/or said morphology Y11 is calculated according to the mathematical law Y11=β.sub.0+Σ.sub.j=1.sup.nβ.sub.jX′.sub.j, where X′.sub.j(j∈[1;n]) is the first derivative of the absorption X.sub.j, normalised by an SNV pre-processing, for the wavenumber range [1,800 cm−1; 900 cm−1] and the weighting coefficients β.sub.0 and β.sub.j(j∈[1;n]) are constants; and/or said osmolarity Y12 is calculated according to the mathematical law Y12=β.sub.0+Σ.sub.j=1.sup.nβ.sub.jX′.sub.j, where X′.sub.j(j∈[1;n]) is the first derivative of the absorption X.sub.j, normalised by an SNV pre-processing, for the wavenumber range [1,700 cm−1; 910 cm−1] and the weighting coefficients β.sub.0 and β.sub.j(j∈[1;n]) are constants; and/or said GSH level Y13 is calculated according to the mathematical law Y13=β.sub.0+Σ.sub.j=1.sup.nβ.sub.jX′.sub.j, where X′.sub.j(j∈[1;n]) is the first derivative of the absorption X.sub.j, normalised by an SNV pre-processing, for the wavenumber range [1,800 cm−1; 900 cm−1] and the weighting coefficients β.sub.0 and β.sub.j(j∈[1;n]) are constants; and/or said non-return rate at 56 days Y14 is calculated according to the mathematical law Y14=β.sub.0+Σ.sub.j=1.sup.nβ.sub.jX′.sub.j, where X′.sub.j(j∈[1;n]) is the first derivative of the absorption X.sub.j, normalised by an SNV pre-processing, for the wavenumber range [1,800 cm−1; 900 cm−1] and the weighting coefficients β.sub.0 and β.sub.j(j∈[1;n]) are constants; and/or said non-return rate at 90 days Y15 is calculated according to the mathematical law Y15=β.sub.0+Σ.sub.j=1.sup.nβ.sub.jX′.sub.j, where X′.sub.j(j∈[1;n]) is the first derivative of the absorption X.sub.j, normalised by an SNV pre-processing, for the wavenumber range [1,800 cm−1; 900 cm−1] and the weighting coefficients β.sub.0 and β.sub.j(j∈[1;n]) are constants; and/or said pregnancy diagnosis Y16 is calculated according to the mathematical law Y16=β.sub.0+Σ.sub.j=1.sup.nβ.sub.jX′.sub.j, where X′.sub.j(j∈[1;n]) is the first derivative of the absorption X.sub.j, normalised by an SNV pre-processing, for the wavenumber range [1,800 cm−1; 900 cm−1] and the weighting coefficients β.sub.0 and β.sub.j(j∈[1;n]) are constants.

5. The method according to claim 1, wherein during the step of measuring, at least two, absorption spectra of at least one sample of said semen are measured and wherein said step of determining a value of the first derivative of the absorption X′.sub.j comprises a step of performing an average of said measured spectra from which said value of the first derivative of the absorption X′.sub.j is determined.

6. The method according to claim 1, which further comprises a step of comparing said at least one calculated parameter with a predetermined threshold specific to said calculated parameter, allowing validating said semen for breeding purposes where said calculated parameter is higher than or equal to said predetermined threshold specific to said parameter or allowing rejecting said semen where said calculated parameter is lower than said predetermined threshold specific to said calculated parameter.

7. The method according to claim 1, wherein the determination of the quality of said fresh semen is obtained between 30 seconds and 5 minutes.

8. The method according to claim 1, wherein the determination of the quality of said frozen semen is obtained between 30 minutes and 90 minutes.

9. The method according to claim 8, which further comprises a step of manufacturing straws for breeding from validated semen.

10. A straw for artificial insemination of semen from a quality animal obtained by implementing the method according to claim 1.

11. A method for using a computer for implementation of the method according to claim 1.

12. A non-transitory software for implementation of the method according to claim 1.

Description

7. DETAILED DESCRIPTION OF THE FIGURES

[0130] FIG. 1 shows a PLS regression model for the parameter of the concentration, made with the MIR spectra of the fresh semen with an SNVD1 pre-processing and cross-validation.

[0131] The concentration Y1 is calculated according to the mathematical law Y1=β.sub.0+Σ.sub.j=1.sup.nβ.sub.jX′.sub.j, where X′.sub.j(j∈[1;n]) is the first derivative of the absorption X.sub.j normalised by an SNV pre-processing for the wavenumber range [1,800 cm.sup.−1; 900 cm.sup.−1] and the weighting coefficients β.sub.0 and β.sub.j(j∈[1;n]) are constants.

[0132] The used balanced matrix consists of 1,392 ejaculates. The obtained R.sup.2 is 0.75 and the RPD is 1.9. These performances allow predicting the concentration of spermatozoa in the sample accurately, quickly and reproducibly from the MIR spectrum.

[0133] The reference values of the concentration are indicated on the abscissa axis and represented by the triangle points (series .box-tangle-solidup. val).

[0134] The predicted values of the concentration are indicated on the ordinate axis and represented by the round points (series .circle-solid. cal).

[0135] The concentration values are expressed with a −10 factor.

[0136] FIG. 2 shows a PLS regression model for the mobility parameter, made with the MIR spectra of the fresh semen with an SNVD1 pre-processing and cross-validation.

[0137] The mobility Y2 is calculated according to the mathematical law Y2=β.sub.0+Σ.sub.j=1.sup.nβ.sub.jX′.sub.j, where X′.sub.j(j∈[1;n]) is the first derivative of the absorption X.sub.j normalised by an SNV pre-processing, for the wavenumber range [1,800 cm−1; 900 cm−1] and the weighting coefficients β.sub.0 and β.sub.j(j∈[1;n]) are constants.

[0138] The used balanced matrix consists of 1,392 ejaculates. The obtained R.sup.2 is 0.63 and the RPD is 1.4 for the freshly measured mobility, and 0.44 and 1.2, respectively, for mobility measured after freezing the semen. These performances allow having an estimate of the fresh mobility of the spermatozoa in the sample, the prediction error approaching half the original error, quickly and reproducibly from the MIR spectrum. The prediction of mobility after freezing features lower and insufficient calibration performances.

[0139] The reference values of the mobility are indicated on the abscissa axis and represented by the triangle points (series .box-tangle-solidup. val).

[0140] The predicted values of the mobility are indicated on the ordinate axis and represented by the round points (series .circle-solid. cal).

[0141] The values of the mobility are expressed with a −10 factor.

[0142] FIG. 3 shows a PLS regression model for the parameter of the progressive spermatozoa rate, made with the MIR spectra of the fresh semen with an SNVD1 pre-processing and cross-validation.

[0143] The progressive spermatozoa rate Y3 is calculated according to the mathematical law Y3=β.sub.0+Σ.sub.j=1.sup.nβ.sub.jX′.sub.j, where X′.sub.j(j∈[1;n]) is the first derivative of the absorption X.sub.j normalised by an SNV pre-processing, for the wavenumber range [1,800 cm−1; 900 cm−1], and the weighting coefficients β.sub.0 and β.sub.j(j∈[1;n]) are constants.

[0144] The used balanced matrix consists of 1,392 ejaculates. The obtained R.sup.2 is 0.57 and the RPD is 1.4 for the freshly measured progressive mobility, and 0.35 and 1.1, respectively, for the progressive mobility measured after freezing the semen. These performances allow having an estimate of the fresh mobility of the spermatozoa in the sample, the prediction error approaching half the original error, quickly and reproducibly from the MIR spectrum. The progressive mobility after freezing features lower calibration performances in comparison with the progressive mobility on the fresh semen.

[0145] The reference values of the progressive spermatozoa rate are indicated on the abscissa axis and represented by the triangle points (series .box-tangle-solidup. val).

[0146] The predicted values of the progressive spermatozoa rate are indicated on the ordinate axis and represented by the round points (series .circle-solid. cal).

[0147] The values of the progressive spermatozoa rate are expressed with a −10 factor.

[0148] FIG. 4 shows a PLS regression model for the viability parameter, made with the MIR spectra of the fresh semen with an SNVD1 pre-processing and cross-validation.

[0149] The viability Y4 is calculated according to the mathematical law Y4=β.sub.0+Σ.sub.j=1.sup.nβ.sub.jX′.sub.j, where X′.sub.j(j∈[1;n]) is the first derivative of the absorption X.sub.j, normalised by an SNV pre-processing, for the wavenumber range [1,800 cm−1; 900 cm−1] and the weighting coefficients β.sub.0 and β.sub.j(j∈[1;n]) are constants.

[0150] The used balanced matrix consists of 1,386 ejaculates. The obtained R.sup.2 is 0.6 and the RPD is 1.3. These performances allow validating that the ejaculates actually measured more than 50% viable are predicted in the class with more than 30% viable spermatozoa (acceptable quality threshold).

[0151] The reference values of the viability are indicated on the abscissa axis and represented by the triangle points (series .box-tangle-solidup. val).

[0152] The predicted values of the viability are indicated on the ordinate axis and represented by the round points (series .circle-solid. cal).

[0153] The values of the viability are expressed with a −9 factor.

[0154] FIG. 5 shows a PLS regression model for the parameter of the percentage of spermatozoa with stable phospholipids, made with the MIR spectra of the fresh semen with an SNVD1 pre-processing and cross-validation.

[0155] The stability of the membrane phospholipids, in other words the percentage of live spermatozoa with stable phospholipids Y5 is calculated according to the mathematical law Y5=β.sub.0+Σ.sub.j=1.sup.nβ.sub.jX′.sub.j, where X′.sub.j(j∈[1;n]) is the first derivative of the absorption X.sub.j, normalised by an SNV pre-processing, for the wavenumber range [1,800 cm−1; 900 cm−1] and the weighting coefficients β.sub.0 and β.sub.j(j∈[1;n]) are constants.

[0156] The used balanced matrix consists of 1,386 ejaculates. The obtained R.sup.2 is 0.6 and the RPD is 1.3. These performances allow validating that the ejaculates actually measured more than 50% viable with a good organisation of membrane phospholipids are predicted in the class with more than 25% of corresponding spermatozoa (acceptable quality threshold).

[0157] The reference values of the stability of the membrane phospholipids are indicated on the abscissa axis and represented by the triangle points (series .box-tangle-solidup. val).

[0158] The predicted values of the stability of the membrane phospholipids are indicated on the ordinate axis and represented by the round points (series .circle-solid. cal).

[0159] The values of the stability of the membrane phospholipids are expressed with a −8 factor.

[0160] FIG. 6 shows a PLS regression model for the parameter of the mitochondrial potential, i.e. for the percentage of spermatozoa with polarised mitochondria, made with the MIR spectra of the fresh semen with an SNVD1 pre-processing and cross-validation.

[0161] The mitochondrial potential Y6 is calculated according to the mathematical law Y6=β.sub.0+Σ.sub.j=1.sup.nβ.sub.jX′.sub.j, where X′.sub.j(j∈[1;n]) is the first derivative of the absorption X.sub.j, normalised by an SNV pre-processing, for the wavenumber range [1,800 cm−1; 900 cm−1] and the weighting coefficients β.sub.0 and β.sub.j(j∈[1;n]) are constants.

[0162] The used unbalanced matrix consists of 1,386 ejaculates. The obtained R.sup.2 is 0.46 and the RPD is 1.1. The ejaculates actually measured more than 45% spermatozoa with well-polarised mitochondria are predicted in the class with more than 30% corresponding spermatozoa (acceptable quality threshold).

[0163] The reference values of the mitochondrial potential are indicated on the abscissa axis and represented by the triangle points (series .box-tangle-solidup. val).

[0164] The predicted values of the mitochondrial potential are indicated on the ordinate axis and represented by the round points (series .circle-solid. cal).

[0165] The values of the mitochondrial potential are expressed with a −7 factor.

[0166] FIG. 7 shows a PLS regression model for the parameter of the percentage of spermatozoa with peroxidised lipids, made with the MIR spectra of the fresh semen with an SNVD1 pre-processing and cross-validation.

[0167] The percentage of spermatozoa with peroxidised lipids Y7 is calculated according to the mathematical law Y7=β.sub.0+Σ.sub.j=1.sup.nβ.sub.jX′.sub.j, where X′.sub.j(j∈[1;n]) is the first derivative of the absorption X.sub.j, normalised by an SNV pre-processing, for the wavenumber range [1,800 cm−1; 900 cm−1] and the weighting coefficients β.sub.0 and β.sub.j(j∈[1;n]) are constants.

[0168] A total matrix of 1,386 ejaculates has been used. The obtained R.sup.2 is 0.53 and the RPD is 1.2.

[0169] The reference values of the percentage of spermatozoa with peroxidised lipids are indicated on the abscissa axis and represented by the triangle points (series .box-tangle-solidup. val).

[0170] The predicted values of the percentage of spermatozoa with peroxidised lipids are indicated on the ordinate axis and represented by the round points (series .circle-solid. cal).

[0171] The values of the percentage of spermatozoa with peroxidised lipids are expressed with a −7 factor.

[0172] FIG. 8 shows a PLS regression model for the parameter of the percentage of spermatozoa with an intact acrosome, made with the MIR spectra of the fresh semen with an SNVD1 pre-processing and cross-validation.

[0173] The percentage of spermatozoa with an intact acrosome Y8 is calculated according to the mathematical law Y8 β.sub.0+Σ.sub.j=1.sup.nβ.sub.jX′.sub.j, where X′.sub.j(j∈[1;n]) is the first derivative of the absorption X.sub.j, normalised by an SNV pre-processing, for the wavenumber range [1,800 cm−1; 900 cm−1] and the weighting coefficients β.sub.0 and β.sub.j(j∈[1;n]) are constants.

[0174] The used unbalanced matrix consists of 1,386 ejaculates. The obtained R.sup.2 is 0.39 and the RPD is 1.2. The ejaculates actually measured more than 75% of spermatozoa with an intact acrosome are predicted in the class with more than 60% of corresponding spermatozoa (acceptable quality threshold).

[0175] The reference values of the percentage of spermatozoa with an intact acrosome are indicated on the abscissa axis and represented by the triangle points (series .box-tangle-solidup. val).

[0176] The predicted values of the percentage of spermatozoa with an intact acrosome are indicated on the ordinate axis and represented by the round points (series .circle-solid. cal).

[0177] The values of the percentage of spermatozoa with an intact acrosome are expressed with a −5 factor.

[0178] FIG. 9 shows a PLS regression model for the parameter of the total antioxidant capacity TAC, made with the MIR spectra of the fresh semen with an SNVD1 pre-processing and cross-validation.

[0179] The total antioxidant capacity TAC Y9 is calculated according to the mathematical law Y9=β.sub.0+Σ.sub.j=1.sup.nβ.sub.jX′.sub.j, where X′.sub.j(j∈[1;n]) is the first derivative of the absorption X.sub.j, normalised by an SNV pre-processing, for the wavenumber range [1,700 cm−1; 910 cm−1] and the weighting coefficients β.sub.0 and β.sub.j(j∈[1;n]) are constants.

[0180] The used unbalanced matrix consists of 59 ejaculates. The obtained R.sup.2 is 0.83 and the RPD is 1.1.

[0181] The reference values of the total antioxidant capacity TAC are indicated on the abscissa axis and represented by the triangle points (series .box-tangle-solidup. val).

[0182] The predicted values of the total antioxidant capacity TAC are indicated on the ordinate axis and represented by the round points (series .circle-solid. cal).

[0183] The values of the total antioxidant capacity TAC are expressed with a −5 factor.

[0184] FIG. 10 is a table listing the R.sup.2s and the RPDs obtained for the parameter of the fatty acid composition, assayed in the spermatozoa, made with the MIR spectra of the fresh semen with an SNVD1 pre-processing and cross-validation, for Belgian Blue and Holstein bulls.

[0185] The composition of the spermatozoa for the different measured fatty acids Y10 is calculated according to the mathematical law Y10=β.sub.0+Σ.sub.j=1.sup.nβ.sub.jX′.sub.j, where X′.sub.j(j∈[1;n]) is the first derivative of the absorption X.sub.j, normalised by an SNV pre-processing, for the wavenumber range [1,800 cm−1; 900 cm−1] and for the wavenumber range [3,000 cm−1; 2,700 cm−1] and the weighting coefficients β.sub.0 and β.sub.j(j∈[1;n]) are constants.

[0186] The used unbalanced matrix consists of 86 Blue Belgian ejaculates and 23 Holstein ejaculates. These performances generally allow predicting, inter alia, the presence of n−3 type fatty acids in the sample accurately and quickly from the MIR spectrum, more advantageously for the most unsaturated ones among them.

[0187] FIG. 11 shows a PLS regression model for the parameter of the percentage of spermatozoa having a normal morphology, made with the MIR spectra of the fresh semen with an SNVD1 pre-processing and cross-validation.

[0188] The percentage of spermatozoa having a normal morphology Y11 is calculated according to the mathematical law Y11=β.sub.0+Σ.sub.j=1.sup.nβ.sub.jX′.sub.j, where X′.sub.j(j∈[1;n]) is the first derivative of the absorption X.sub.j, normalised by an SNV pre-processing, for the wavenumber range [1,800 cm−1; 900 cm−1] and the weighting coefficients β.sub.0 and β.sub.j(j∈[1;n]) are constants.

[0189] The used unbalanced matrix consists of 1,392 ejaculates. The obtained R.sup.2 is 0.45 and the RPD is 1.3. The ejaculates actually measured more than 95% spermatozoa having a normal morphology are predicted in the class with more than 80% corresponding spermatozoa (acceptable quality threshold).

[0190] The reference values of the percentage of spermatozoa having a normal morphology are indicated on the abscissa axis and represented by the triangle points (series .box-tangle-solidup. val).

[0191] The predicted values of the percentage of spermatozoa having a normal morphology are indicated on the ordinate axis and represented by the round points (series .circle-solid. cal).

[0192] The values of the percentage of spermatozoa having a normal morphology are expressed with a −7 factor.

[0193] FIG. 12 shows a PLS regression model for the parameter of the osmolarity, made with the MIR spectra of the fresh semen with an SNVD1 pre-processing and cross-validation.

[0194] The osmolarity Y12 is calculated according to the mathematical law Y12=β.sub.0+Σ.sub.j=1.sup.nβ.sub.jX′.sub.j, where X′.sub.j(j∈[1;n]) is the first derivative of the absorption X.sub.j, normalised by an SNV pre-processing, for the wavenumber range [1,700 cm−1; 910 cm−1] and the weighting coefficients β.sub.0 and β.sub.j(j∈[1;n]) are constants.

[0195] The used unbalanced matrix consists of 56 ejaculates. The obtained R.sup.2 is 0.88 and the RPD is 1.4. These performances allow having an estimate of the osmolarity of the sample (the prediction error approaches half the original error) quickly from the MIR spectrum.

[0196] The reference values of the osmolarity are indicated on the abscissa axis and represented by the triangle points (series .box-tangle-solidup. val).

[0197] The predicted values of the osmolarity are indicated on the ordinate axis and represented by the round points (series .circle-solid. cal).

[0198] The values of the osmolarity are expressed with a −4 factor.

[0199] FIG. 13 shows a PLS regression model for the parameter of the glutathione GSH level, made with the MIR spectra of the fresh semen with an SNVD1 pre-processing and cross-validation.

[0200] The glutathione GSH level Y13 is calculated according to the mathematical law Y13=β.sub.0+Σ.sub.j=1.sup.nβ.sub.jX′.sub.j, where X′.sub.j(j∈[1;n]) is the first derivative of the absorption X.sub.j, normalised by an SNV pre-processing, for the wavenumber range [1,800 cm−1; 900 cm−1] and the weighting coefficients β.sub.0 and β.sub.j(j∈[1;n]) are constants.

[0201] The used unbalanced matrix consists of 56 ejaculates. The obtained R.sup.2 is 0.78 and the RPD is 1.04.

[0202] The reference values of the glutathione level are indicated on the abscissa axis and represented by the triangle points (series .box-tangle-solidup. val).

[0203] The predicted values of the glutathione level are indicated on the ordinate axis and represented by the round points (series .circle-solid. cal).

[0204] The values of the glutathione level are expressed with a −5 factor.

[0205] FIG. 14 shows a PLS regression model for the parameter of the non-return rate at 56 days, made with the MIR spectra of the fresh semen with an SNVD1 pre-processing and cross-validation.

[0206] The non-return rate at 56 days Y14 is calculated according to the mathematical law 14=β.sub.0+Σ.sub.j=1.sup.nβ.sub.jX′.sub.j, where X′.sub.j(j∈[1;n]) is the first derivative of the absorption X.sub.j, normalised by an SNV pre-processing, for the wavenumber range [1,800 cm−1; 900 cm−1] and the weighting coefficients β.sub.0 and β.sub.j(j∈[1;n]) are constants.

[0207] The unbalanced matrix made on the basis of the spectra acquired on the fresh semen (before freezing) consists of 96 ejaculates. The obtained R.sup.2 is 0.86 and the RPD is 1.6. These performances allow having an estimate of the TNR56 (the prediction error is equal to half the original error) quickly and reproducibly from the MIR spectrum.

[0208] The reference values of the non-return rate at 56 days are indicated on the abscissa axis and represented by the triangle points (series .diamond-solid. val).

[0209] The predicted values of the non-return rate at 56 days are indicated on the ordinate axis and represented by the round points (series .circle-solid. cal).

[0210] The values of the non-return rate at 56 days are expressed with a −7 factor.

[0211] Furthermore, the unbalanced matrix made on the basis of the spectra acquired on the frozen semen consists of 162 ejaculates. The obtained R.sup.2 is 0.53 and the RPD is 1.2. These predictions after freezing the semen are less accurate yet allow identifying the very good and the very bad semen batches.

[0212] FIG. 15 shows a PLS regression model for the parameter of the non-return rate at 90 days, made with the MIR spectra of the fresh semen with an SNVD1 pre-processing and cross-validation.

[0213] The non-return rate at 90 days Y15 is calculated according to the mathematical law Y15=β.sub.0+Σ.sub.j=1.sup.nβ.sub.jX′.sub.j, where X′.sub.j(j∈[1;n]) is the first derivative of the absorption X.sub.j, normalised by an SNV pre-processing, for the wavenumber range [1,800 cm−1; 900 cm−1] and the weighting coefficients β.sub.0 and β.sub.j(j∈[1;n]) are constants.

[0214] The unbalanced matrix made on the basis of the spectra acquired on the fresh semen (before freezing) consists of 67 ejaculates. The obtained R.sup.2 is 0.82 and the RPD is 1.5. These performances allow having an estimate of the TNR at 90 days from the ejaculate (the prediction error is equal to half the original error) quickly and reproducibly from the MIR spectrum.

[0215] The reference values of the non-return rate at 90 days are indicated on the abscissa axis and represented by the triangle points (series .box-tangle-solidup. val).

[0216] The predicted values of the non-return rate at 90 days are indicated on the ordinate axis and represented by the round points (series .circle-solid. cal).

[0217] The values of the non-return rate at 90 days are expressed with a −5 factor.

[0218] Furthermore, the unbalanced matrix made on the basis of spectra acquired on frozen semen consists of 99 ejaculates. The obtained R.sup.2 is 0.86 and the RPD is 1.3. These predictions after freezing the semen are less accurate yet allow identifying the very good and the very bad semen batches.

[0219] FIG. 16 shows a PLS regression model for the parameter of the pregnancy diagnosis, made with the MIR spectra of the fresh semen with an SNVD1 pre-processing and cross-validation.

[0220] the pregnancy diagnosis Y16 is calculated according to the mathematical law Y16=β.sub.0+Σ.sub.j=1.sup.nβ.sub.jX′.sub.j, where X′.sub.j(j∈[1;n]) is the first derivative of the absorption X.sub.j, preferably normalised by an SNV pre-processing, for the wavenumber range [1,800 cm−1; 900 cm−1] and the weighting coefficients β.sub.0 and β.sub.j(j∈[1;n]) are constants.

[0221] The unbalanced matrix made on the basis of the spectra acquired on the fresh semen (before freezing) consists of 49 ejaculates. The obtained R.sup.2 is 0.79 and the RPD is 1.6. These performances allow having an estimate of the pregnancy diagnosis of the semen (the prediction error is equal to half the original error) in a rapid and reproducible manner from the MIR spectrum on the fresh semen.

[0222] The reference values for the pregnancy diagnosis are indicated on the abscissa axis and represented by the triangle points (series .box-tangle-solidup. val).

[0223] The predicted values of the pregnancy diagnosis are indicated on the ordinate axis and represented by the round points (series .circle-solid. cal).

[0224] The values of the pregnancy diagnosis are expressed with a −5 factor.

[0225] Furthermore, the unbalanced matrix made on the basis of spectra acquired on the frozen semen consists of 81 ejaculates. The obtained R.sup.2 is 0.84 and the RPD is 1.8. These performances allow having an estimate of the pregnancy diagnosis (the prediction error is equal to half the original error) in a rapid and reproducible manner from the MIR spectrum measured on the frozen semen.

[0226] It should be understood that the present invention is in no way limited to the above-described embodiments and that many modifications could be made thereto without departing from the scope of the appended claims.