seminal MOTility Imaging and MOdeling for automatized fertility assessment – MOTIMO

The goal of the MOTIMO project is to develop the concepts necessary to the design of automated fertility tests of animal semen samples. The turbulence of the semen liquid induced by spermatozoa motion, the so-called massal motility can be observed through a microscope. It has been established that massal motility is an excellent fertility criterion for the semen sample, while the individual spermatozoon motility is not. The project aims at developing imaging and scoring techniques which will allow for an automated selection of the fertile samples and the rejection of the other ones, through the observation of this motility. These techniques will then be improved through the use of adequate models. The modeling of the collective mechanisms which lead to this massal motility will provides clues to understand why spermatozoon individual motility is not well correlated with massal motility. The whole project, which should lead to an industrializable process, will be supervised by a industrial leader in the sector of Artificial Insemination (AI), the company IMV-Technologies. The consortium gathers groups covering the complete chain going from biological experimentation and imaging to statistical analysis and modeling. The ‘Laboratoire de Physiologie de la Reproduction et du Comportement’ of INRA-Tours will be in charge of the microscope measurements and of the process validation through the realization of an AI campaign with scored samples. The ‘Institut de Mécanique des Fluides de Toulouse (IMFT)’ will perform complementary experiments and will be in charge of the processing of the collected images, through the dual use of ‘Particle Imaging Velocimetry’ (PIV) and ‘Optical Flow Reconstruction’ (OFR) techniques. These techniques will lead to a flow reconstruction. The ‘Laboratoire d’Informatique Signaux et Systèmes (I3S)’ of Sophia-Antipolis will be in charge of the learning-scoring step, which, from the reconstructed flows, will provide an automatic score of the considered sample. Finally, the ‘Institut de Mathématiques de Toulouse (IMT)’ will take care of the modeling phase, which will lead to a detailed understanding and a better flow reconstruction. Several types of microscopic recording will be done. First, phase contrast microscopy realized at INRA will lead to the direct observation of the waves within semen. Another technique also developed at INRA consists in observing the motion of passive beads mixed with the sample. Finally, individual motility measurements will be also made at INRA. At IMFT, complementary experiments will allow the simultaneous observation of spermatozoa and of the beads. In parallel, the IMFT will be in charge of the image processing phase through PIV and also, in collaboration with IMT and I3S, through OFR. The I3S will develop the scoring and learning techniques thanks to subjective scores delivered by AI professionals. The IMT, in partnership with IMFT will establish and implement models. Two kinds of models will be designed and developed. A first microscopic model (‘Individual-Based Model’ or IBM) will be based on individual spermatozoa motility and their hydrodynamic interaction. The other one, of macroscopic nature, will be obtained from the first one through kinetic theory techniques and will represent the ensemble fluid motion. After validation and calibration, these models will be used to improve the OFR. This multi-disciplinary project will lead first to a better scientific insight into fertility mechanisms on the one hand, and to a process the industrial capability of which will be solidly established.