Research On The Prediction Of The Fingerprint Affiliated To Amino Acids Metabolism Of Embryo Culture Media For The Potential Of Embyro Implantation

Objective:In recent years,assisted reproductive technology turns to be the major method for treating of infertility.High multiple pregnancy rate does harm to the health of maternal and infant health seriously.At the same time,single embryo transplant becomes the most effective method to reduce the rate of multiple pregnancy.Thus,it is especially important to establish a stable and reliable method for the selection of embryo.In order to provide the basis for the selection of high-quality embryo,improve the pregnancy rate and reduce the multiple pregnancy rate,this study explored the relation between the embryo implantation and amino acid metabolism of culture medium.In addition,this research also analyzed the fingerprint of amino acid metabolism.And the prediction model of embryo quality was established based on the fingerprint of amino acid metabolism with the prediction model of embryo quality which was established based on the fingerprint of amino acid metabolism.Methods:The data of 98 embryo culture media during cleavage of 49 couples received IVF-ET/ICSI-ET treatment in our hospital from2015 December to 2017 February were retrospectively analyzed.This study divided the 98 embryo culture media into experimental group(gemellary pregnancy)and negative control group(non-pregnancy).The quantity of embryo culture media of experimental group and negative control group were 52 and 46,respectively.At the same time,this research also collected 48 blank cleavage medium as the blank control group.Next,this research collected the nutrient solution of embryo from the three groups and preserved them in liquid nitrogen.At last,the ACCQ reagents were used for the derivatization reaction of amino acids.In addition,the substance produced above were detected by high performance liquid chromatograp.This research established the prediction model and explored the biomarker of the model and the value of the prediction through the principal component analysis,partial least squares analysis and discriminant analysis.And this study also explored the biomarker of the model and the value of the prediction.Results:(1)This study got the PCA model through the principal component analysis of the unsupervised algorithm.The rate of prediction of embryo implantation from this research(63.1%)was better than that of the related data(30%-40%)from our national auxiliary reproductive data reporting system.(2)We also got the PLA-DA model through the partial least squares analysis of the supervise algorithm.The intercept of R~2Y and Q~2 were 0.0131 and-0.176,respectively.The embryo implantation rate of this model was 72.4%.In addition,this research indicated that the prediction effect of the PLA-DA was better than that of PCA.(3)The discrimination model(prediction model-1)according to discriminant analysis revealed that the prediction rate of the experimental group,negative control group and blank control group were 80.8%,95.7%and 100%,respectively.This indicated that the prediction effect was basically satisfied.However,the prediction rate of the experimental group was not reached the optimal state and we used more data paramaters.(4)The biomarkers of this research according to the analysis of the PLS-DA model and the independent-samples T test were H-SER,S-SER and S-ALA.This research got the prediction model-2 according to the data analyzed from the biomarker.The cross validation of this model indicated that the prediction rate of the experimental group,negative control group and blank control group were 82.7%,95.7%and 100%,respectively.In addition,the involved parameters of prediction model-2 was less than that of prediction model-1.Then this research identified the prediction model-2as the satisfying model with better prediction effect and prediction model-2 is more suitable for clinical application.Conclusions:The fingerprint of amino acid metabolism of embryo culture medium from cleavage stage could predict the implantation potential.The prediction rate of biomarker-discrimination model(92.5%)was better than that of original-discrimination model,PCA model and PLS-DA model.At the same time,the prediction rate of the experimental group was82.7%.The biomarker(H-SER,S-SER and S-ALA)did contribute to establish the prediction model of implantation potency.Additionally,the biomarker did contribute to select the high quality embryo with better ability of implantation.