Optimization and anomaly judgment of NIPT detection based on multivariate statistical model and machine learning

Ziqi Zhao; Yangsai Zhou; Qianziyi Guo

doi:10.54097/aqmjsx77

Authors

Ziqi Zhao
Yangsai Zhou
Qianziyi Guo

DOI:

https://doi.org/10.54097/aqmjsx77

Keywords:

Non-invasive prenatal testing, Multiple linear regression, Cluster analysis, Spline interpolation, Error analysis, Machine learning

Abstract

In this study, focusing on non-invasive prenatal detection technology, using the individual characteristics and sequencing data of pregnant women, a multivariate statistical model and machine learning algorithm were constructed to systematically analyze the relationship between fetal sex chromosome concentration changes and the optimal detection time. Firstly, through Spearman rank correlation analysis and multiple linear regression model, the significant effects of gestational age and BMI on Y chromosome concentration were revealed. Secondly, the pregnant women were divided into three groups according to BMI by K-means clustering, and the optimal time points for each group were determined by linear fitting, which were 12 weeks and 2 days (low BMI group), 16 weeks and 4 days (medium BMI group) and 18 weeks and 5 days (high BMI group). The thin plate spline interpolation model was further introduced to capture the nonlinear interaction effect between age, height and Y concentration, and the X chromosome concentration was identified as the main source of technical error. The results show that the NIPT time-point recommendation strategy based on BMI grouping and multivariate modeling can significantly improve the reliability of detection and provide a theoretical basis for clinical personalized detection.

Downloads

Download data is not yet available.

References

[1] Lo, Y. M., et al. (2021). Noninvasive prenatal testing: a review of the current state of the science. Annual Review of Genomics and Human Genetics, 22, 1-25.

[2] Chen, L., et al. (2021). Fetal fraction estimation in maternal plasma and its association with gestational age and maternal weight. Prenatal Diagnosis, 41(4), 456-463.

[3] Liu, Y., et al. (2022). Impact of maternal BMI on fetal DNA fraction in noninvasive prenatal testing. Journal of Translational Medicine, 20(1), 1-10.

[4] Sun, K., et al. (2020). Sequencing depth and coverage uniformity impacts on NIPT performance. Clinical Chemistry, 66(8), 1072-1081.

[5] Wu, H., et al. (2021). GC bias correction in cell-free DNA sequencing for aneuploidy detection. Bioinformatics, 37(16), 2389-2395.

[6] Liao, C., et al. (2022). Biological determinants of fetal fraction variation in maternal plasma. Human Genetics, 141(5), 987-996.

[7] Zhang, H., et al. (2021). Optimization of NIPT timing through multivariate regression modeling. Clinical Genetics, 99(3), 321-329.

[8] Chen, S., et al. (2022). Longitudinal study of fetal DNA dynamics across gestation. Prenatal Diagnosis, 42(1), 45-52.

[9] Liu, J., et al. (2023). Machine learning for predicting fetal fraction from maternal characteristics. IEEE Journal of Biomedical and Health Informatics, 27(1), 210-218.

[10] Wang, Q., et al. (2023). Machine learning-based prediction of fetal aneuploidy using cell-free DNA sequencing data. Bioinformatics, 39(5), 1-9.

[11] Xu, R., et al. (2022). Nonlinear relationship between maternal characteristics and fetal DNA concentration. PLOS ONE, 17(6), e0268888.

[12] Li, M., et al. (2023). A clustering-based approach for personalized NIPT scheduling. Medical & Biological Engineering & Computing, 61(4), 987-996.

[13] Zhou, Y., et al. (2023). Multivariate spline interpolation for biomedical data analysis. Computational Statistics & Data Analysis, 179, 107-120.

[14] Guo, L., et al. (2021). Error source analysis in noninvasive prenatal testing: a systematic review. Clinical Chemistry and Laboratory Medicine, 59(8), 1321-1330.

[15] Pan, Y., et al. (2023). Spearman correlation analysis in nonparametric medical data modeling. Journal of Applied Statistics, 50(5), 1123-1135.