Classification of sperm cells according to their chromosomic content using a neural network trained with a genetic algorithm

A priori determination of the sex of a human individual before gestation is a desirable goal in some cases. To achieve this, it is necessary to perform the separation of sperm cells containing either X or Y chromosomes. As is well known, male sex depends on the presence of chromosome Y. Once this separation is achieved in principle, we require to determine, with a high degree of accuracy, whether the sperm cells of interest contain the desired X or Y chromosomes. If we are able to obtain certain simple measurements regarding the sperm cells under consideration we will be able to control the fertilization process reliably. In this paper we report a method which allows for non-invasive verification of the characteristics of the separated sperm. We determined a set of easily measurable characteristics. From a sample drawn from previously cropped sperm we trained a neural network with a genetic algorithm. The trained network was able to perform a posteriori classification with an error much smaller than 1%. This percentage of efficiency is better than the ones reported in centers of assisted fecundation.