Receive a weekly summary and discussion of the top papers of the week by leading researchers in the field.

In Reproductive sciences (Thousand Oaks, Calif.)

The aim of this study is to explore the relationship between mitochondrial DNA (mtDNA) copy number and embryo implantation potential in in vitro fertilization (IVF). A retrospective study of 319 blastocysts from patients undergoing preimplantation genetic testing (PGT) at Reproductive Medicine Center in Tongji Hospital from January 2016 to February 2018 was conducted. We used multiple annealing- and looping-based amplification cycles (MALBAC) technology to amplify the genetic materials from the trophectoderm cells of blastocysts, and next-generation sequencing (NGS) technology to test mitochondrial DNA copy number. Box-Cox transformation was introduced to eliminate the skewness distribution of mtDNA copy number, and the transformed data were defined as adjusted mtDNA. Subsequently, associations between adjusted mtDNA and the clinical characteristics of patients were assessed by univariate analysis and multiple linear regression. In addition, Gaussian Naive Bayes classifier was also used to predict pregnancy outcomes. We observed that only antral follicle count (AFC) was significantly associated with adjusted mtDNA without the influence of multicollinearity. What's more, the distribution of the adjusted mtDNA of blastocysts resulting in live birth was more concentrated than that of others. The area under the curve (AUC) of the prediction model that combined adjusted mtDNA with other clinical characteristics of patients was up to 0.81, higher than that excluded adjusted mtDNA. Among patient clinical characteristics, AFC was significantly associated with adjusted mtDNA. Mitochondrial DNA copy number may help to optimize the pregnancy outcome prediction in IVF.

Zhu Lixia, Li Jingjing, Wang Meng, Fang Zishui, Zheng Fangqin, Li Zhou, Jin Lei

2021-Jan-05

Assisted reproductive technology, Box-Cox transformation, Machine learning, Mitochondrial DNA copy number, Pregnancy outcome prediction