Related Experiment Video
Updated: Aug 24, 2025
Semiconductor Sequencing for Preimplantation Genetic Testing for Aneuploidy
Published on: August 25, 2019
Application of machine learning to predict aneuploidy and mosaicism in embryos from in vitro fertilization cycles
José A Ortiz1, Ruth Morales1, Belén Lledó1
1Department of Molecular Biology and Genetics, Instituto Bernabeu, Alicante, Spain (Drs Ortiz, Morales, Lledó, and García-Hernández, and MSc Cascales).
Background:
The factors associated with embryo aneuploidy have been extensively studied. Mostly maternal age and to a lesser extent male factor and ovarian stimulation have been related to the occurrence of chromosomal alterations in the embryo. On the other hand, the main factors that may increase the incidence of embryo mosaicism have not yet been established.
Objective:
This study aimed to establish a machine learning model that would allow prediction of aneuploidies and mosaicism in embryos conceived via in vitro fertilization, and thus help to determine which variables are associated with these chromosomal alterations.
Study Design:
The study design was observational and retrospective. A total of 6989 embryos from 2476 cycles of preimplantation genetic testing for aneuploidies were included (January 2013 to December 2020). The trophoectoderm biopsies on day-5, -6, or -7 blastocysts were analyzed by preimplantation genetic testing for aneuploidies (PGT-A). The different maternal, paternal, couple, embryo, and in vitro fertilization cycle characteristics were recorded in a database (22 predictor variables) from which predictive models of embryo aneuploidy and mosaicism were developed; 16 different unsupervised classification machine learning algorithms were used to establish the predictive models.
Results:
Two different predictive models were performed: one for aneuploidy and the other for mosaicism. The predictor variable was of multiclass type because it included the segmental- and whole-chromosome alteration categories. The best predicting models for both aneuploidies and mosaicism were those obtained from the Random Forest algorithm. The area under ROC curve (AUC) value was 0.792 for the aneuploidy explanatory model and 0.776 for mosaicism. The most important variable in the final aneuploidy model was maternal age, followed by paternal and maternal karyotype and embryo quality. In the predictive model of mosaicism, the most important variable was the technique used in preimplantation genetic testing for aneuploidies and embryo quality, followed by maternal age and day of biopsy.
Conclusion:
It is possible to predict embryo aneuploidy and mosaicism from certain characteristics of the patients and their embryos.
