Taking advantage of the use of supervised learning methods for characterization of sperm population structure related with freezability in the Iberian red deer.

Using Iberian red deer as a model, this study presents a supervised learning method, the Support Vector Machines (SVM), to characterize sperm population structure related with freezability. Male freezability was assessed by evaluating motility, membrane status and mitochondrial membrane potential of sperm after a freezing-thawing procedure. The SVM model was generated using sperm motility information captured by computer-assisted sperm analysis (CASA) from thawed semen, belonging to six stags with marked differences on their freezability. A total of 1369 sperm tracks were recorded for seven kinematic parameters and assigned to four motility patterns based on them: weak motile, progressive, transitional and hyperactivated-like. Then, these data were split in two sets: the training set, used to train the SVM model, and the testing set, used to examine how the SVM method and three other unsupervised methods, a non-hierarchical, a hierarchical and a multistep clustering procedures, performed the sperm classification into subpopulations. The SVM was revealed as the most accurate method in the characterization of sperm subpopulations, showing all the sperm subpopulations obtained in this way high significant correlations with those sperm parameters used to characterize freezability of males. Given its superiority, the SVM method was used to characterize the sperm motile subpopulations in Iberian red deer. Sperm motile data from frozen-thawed semen belonging to 25 stags were recorded and loaded into the SVM model. The sperm population structure revealed that those males showing poor freezability were characterized by high percentages of sperm with a weak motility pattern. In opposite, males showing good freezability were characterized by higher percentages of sperm with a progressive and hyperactivated-like motility pattern and lower percentages of sperm with a weak motile pattern. We also identified a sperm subpopulation with a transitional motility pattern. This subpopulation increased as the freezability of males improved, and may be used as indicative of overall sperm motility.

[1]  L. Gallego,et al.  Reproductive seasonality in female Iberian red deer (Cervus elaphus hispanicus). , 2002, Theriogenology.

[2]  Nello Cristianini,et al.  An Introduction to Support Vector Machines and Other Kernel-based Learning Methods , 2000 .

[3]  Klaus Nordhausen,et al.  The Elements of Statistical Learning: Data Mining, Inference, and Prediction, Second Edition by Trevor Hastie, Robert Tibshirani, Jerome Friedman , 2009 .

[4]  P. Rousseeuw.  Silhouettes: a graphical aid to the interpretation and validation of cluster analysis , 1987 .

[5]  A. Petrunkina,et al.  Heterogeneous osmotic behaviour in boar sperm populations and its relevance for detection of changes in plasma membrane. , 2000, Reproduction, fertility, and development.

[6]  Wei Wang,et al.  Classification of Mouse Sperm Motility Patterns Using an Automated Multiclass Support Vector Machines Model1 , 2011, Biology of reproduction.

[7]  S. Grunewald,et al.  Deterioration of plasma membrane is associated with activated caspases in human spermatozoa. , 2003, Journal of andrology.

[8]  M. T. Dinis,et al.  Multivariate cluster analysis to study motility activation of Solea senegalensis spermatozoa: a model for marine teleosts. , 2008, Reproduction.

[9]  P. de Paz,et al.  Sperm Subpopulations in Iberian Red Deer Epididymal Sperm and Their Changes Through the Cryopreservation Process1 , 2005, Biology of reproduction.

[10]  W. Holt.  CAN WE PREDICT FERTILITY RATES? MAKING SENSE OF SPERM MOTILITY , 1995 .

[11]  F. Peña,et al.  A three-step statistical procedure to identify sperm kinematic subpopulations in canine ejaculates: changes after cryopreservation. , 2006, Reproduction in domestic animals = Zuchthygiene.

[12]  D Haussler,et al.  Knowledge-based analysis of microarray gene expression data by using support vector machines. , 2000, Proceedings of the National Academy of Sciences of the United States of America.

[13]  William Stafiord Noble,et al.  Support vector machine applications in computational biology , 2004 .

[14]  J. Dorado,et al.  Changes in the structures of motile sperm subpopulations in dog spermatozoa after both cryopreservation and centrifugation on PureSperm(®) gradient. , 2011, Animal reproduction science.

[15]  Bernhard Schölkopf,et al.  Kernel Methods in Computational Biology , 2005 .

[16]  Aixia Guo,et al.  Gene Selection for Cancer Classification using Support Vector Machines , 2014 .

[17]  B. Pérez-Llano,et al.  Four subpopulations of boar spermatozoa defined according to their response to the short hypoosmotic swelling test and acrosome status during incubation at 37 degrees C. , 2003, Theriogenology.

[18]  P. Mermillod,et al.  Assessment of in vitro fertility of deer spermatozoa by heterologous IVF with zona-free bovine oocytes. , 2001, Theriogenology.

[19]  D. Haussler,et al.  Knowledge-based analysis of microarray gene expression , 2000 .

[20]  L. Gallego,et al.  Postmortem assessment of sperm characteristics of the red deer during the breeding season. , 1998, Archives of andrology.

[21]  A. Günzel-Apel,et al.  Focus on Determinants of Male Fertility Determinants of sperm quality and fertility in domestic species , 2007 .

[22]  Shaul Markovitch,et al.  Anytime Learning of Decision Trees , 2007, J. Mach. Learn. Res..

[23]  A. Peña,et al.  Effects of cryopreservation on the motile sperm subpopulations in semen from Asturiana de los Valles bulls. , 2009, Theriogenology.

[24]  Christina S. Leslie,et al.  Fast String Kernels using Inexact Matching for Protein Sequences , 2004, J. Mach. Learn. Res..

[25]  Michalis Vazirgiannis,et al.  Cluster validity methods: part I , 2002, SGMD.

[26]  A. J. Soler,et al.  Functional significance of the sperm head morphometric size and shape for determining freezability in iberian red deer (Cervus elaphus hispanicus) epididymal sperm samples. , 2006, Journal of andrology.

[27]  Bernhard Schölkopf,et al.  Support Vector Machine Applications in Computational Biology , 2004 .

[28]  Jean-Philippe Vert.  A tree kernel to analyze phylog enetic profi les , 2002 .

[29]  P. de Paz,et al.  Post mortem time and season alter subpopulation characteristics of Iberian red deer epididymal sperm. , 2005, Theriogenology.

[30]  HalkidiMaria,et al.  Cluster validity methods , 2002 .

[31]  F. Martínez-Pastor,et al.  Mitochondrial activity and forward scatter vary in necrotic, apoptotic and membrane-intact spermatozoan subpopulations. , 2008, Reproduction, fertility, and development.

[32]  W. Holt,et al.  Fundamental aspects of sperm cryobiology: the importance of species and individual differences. , 2000, Theriogenology.

[33]  R. Harrison,et al.  Capacitation mechanisms, and the role of capacitation as seen in eutherian mammals. , 1996, Reproduction, fertility, and development.

[34]  L. M. Thurston,et al.  Semen cryopreservation: a genetic explanation for species and individual variation? , 2002, Cryo letters.

[35]  L. M. Thurston,et al.  Morphologically distinct sperm subpopulations defined by Fourier shape descriptors in fresh ejaculates correlate with variation in boar semen quality following cryopreservation. , 2001, Journal of andrology.

[36]  E. Cabrita,et al.  Effect of cryopreservation on fish sperm subpopulations. , 2011, Cryobiology.

[37]  Chih-Jen Lin,et al.  LIBSVM: A library for support vector machines , 2011, TIST.

[38]  Jeffrey S. Simonoff,et al.  Tree Induction Vs Logistic Regression: A Learning Curve Analysis , 2001, J. Mach. Learn. Res..

[39]  Robert Tibshirani,et al.  The Elements of Statistical Learning: Data Mining, Inference, and Prediction, 2nd Edition , 2001, Springer Series in Statistics.

[40]  E. Tizado,et al.  Statistical Series: Opportunities and challenges of sperm motility subpopulation analysis. , 2011, Theriogenology.

[41]  A. Valcárcel,et al.  A lineal equation for the classification of progressive and hyperactive spermatozoa. , 1998, Mathematical biosciences.

[42]  Katya Scheinberg,et al.  An Efficient Implementation of an Active Set Method for SVMs , 2006, J. Mach. Learn. Res..

[43]  Jean-Philippe Vert,et al.  A tree kernel to analyse phylogenetic profiles , 2002, ISMB.

[44]  A. García,et al.  Effects of thawing procedure on postthawed in vitro viability and in vivo fertility of red deer epididymal spermatozoa cryopreserved at -196 degrees C. , 2003, Journal of andrology.

[45]  Brian D. Ripley,et al.  Pattern Recognition and Neural Networks , 1996 .

[46]  J. Mesirov,et al.  Molecular classification of cancer: class discovery and class prediction by gene expression monitoring. , 1999, Science.

[47]  Jason Weston,et al.  Mismatch string kernels for discriminative protein classification , 2004, Bioinform..

[48]  J. Gatti,et al.  Hypotonic resistance of boar spermatozoa: sperm subpopulations and relationship with epididymal maturation and fertility. , 2009, Reproduction.

[49]  Jason Weston,et al.  Gene Selection for Cancer Classification using Support Vector Machines , 2002, Machine Learning.

[50]  William Stafford Noble,et al.  Support vector machine , 2013 .

[51]  H. Rodríguez-Martínez,et al.  Kinematic changes during the cryopreservation of boar spermatozoa. , 2005, Journal of andrology.

[52]  Kilian Q. Weinberger,et al.  Distance Metric Learning for Large Margin Nearest Neighbor Classification , 2005, NIPS.

[53]  G. Evans,et al.  Storage of semen and artificial insemination in deer. , 2000, Animal reproduction science.

[54]  Li Liao,et al.  Combining pairwise sequence similarity and support vector machines for remote protein homology detection , 2002, RECOMB '02.

[55]  P. de Paz,et al.  Cryopreservation of Iberian red deer (Cervus elaphus hispanicus) spermatozoa obtained by electroejaculation. , 2009, Theriogenology.

[56]  G. Doncel,et al.  Human sperm subpopulations: relationship between functional quality and protein tyrosine phosphorylation. , 2004, Human reproduction.