Author(s)

JiaJun Tang

Affiliation(s)

School of Automation, Nanjing University of Science and Technology, Nanjing 210000, Jiangsu, China.

Corresponding Author

JiaJun Tang

ABSTRACT

This article focuses on four major issues: correlation analysis of Y chromosome concentration in male fetuses, BMI grouping and determination of optimal detection time points, optimization of time points under multiple factors, and determination of chromosomal abnormalities in female fetuses. Firstly, based on the principle of NIPT technology as the accurate benchmark for results, we analyzed the correlation characteristics between Y chromosome concentration and gestational age, BMI, constructed an adapted relationship model, and verified its significance; Secondly, with the goal of minimizing potential risks, the optimal NIPT timing for male pregnant women with different BMI ranges is determined through reasonable grouping methods, and the impact of detection errors is analyzed; Furthermore, by integrating multiple factors such as height, weight, and age, the time point model is optimized to ensure the reliability of the results; Finally, based on the characteristic of female fetuses without Y chromosome, combined with indicators such as Z value and GC content of X chromosome and chromosomes 21, 18, and 13, an abnormality determination method is established. The research results can provide scientific support for optimizing clinical NIPT testing conditions, improving accuracy, and reducing potential risks for pregnant women, meeting the practical needs of early and accurate detection.This study first preprocessed male fetal data, screened 10-25 week samples, converted gestational weeks into continuous values, and processed duplicate samples and outliers. Use Q-Q chart to test normality and determine the conclusion that Y chromosome concentration, gestational age, and BMI are not normally distributed. Through Spearman rank correlation quantification, it was found that gestational age is strongly positively correlated with Y chromosome concentration, while BMI is moderately negatively correlated with Y chromosome concentration. Construct a quadratic polynomial model for OLS fitting, confirm no severe collinearity through VIF test, and then verify the significance of the model through F-test and t-test. In addition, analysis shows that weight has a significant negative impact on Y chromosome concentration, and this model can reliably reflect the relationship between Y chromosome concentration and gestational age, BMI.Therefore, this study is of great significance. By constructing a multi-model framework that integrates statistical analysis and machine learning, it deepens our understanding of the dynamic patterns of fetal cell-free DNA and promotes cross-disciplinary innovation between bioinformatics and clinical medicine. Meanwhile, this study also has significant practical value by precisely identifying key factors affecting detection accuracy, optimizing the timing of individualized testing, effectively improving the sensitivity and specificity of NIPT, and reducing the risks of missed and incorrect diagnoses, thereby providing data support and a decision-making basis for developing scientific, safe, and efficient prenatal screening strategies in clinical practice. Furthermore, the research findings respond to the urgent national needs of the “Healthy Birth, Healthy China” strategy for a high-precision birth defect prevention and control system, facilitating the shift of prenatal screening from being “experience-driven” to “evidence-driven,” enhancing public health services, and alleviating the medical burden on families and society.

KEYWORDS

Spearson rank correlation coefficient; Kaplan Meier survival analysis; Lasso model; XGBoost model

CITE THIS PAPER

JiaJun Tang. The optimization problem for time-point detection in NIPT based on multi-model fusion. World Journal of Information Technology. 2025, 3(5): 48-55. DOI: https://doi.org/10.61784/wjit3067.

REFERENCES

[1] Suzhou Municipal Hospital team. The influence of maternal age, gestational age, and body mass index on the proportion of fetal free DNA in peripheral blood of pregnant women. Chinese Journal of Prenatal Diagnosis (Electronic Edition), 2017, 9 (3): 1-5.

[2] Lan Zhiheng. Application of DNA sequence distribution ratio method in non-invasive prenatal fetal chromosomal aneuploidy detection. South China University of Technology, 2020.

[3] Obstetrics and Gynecology Team of the Second Affiliated Hospital of Soochow University. Study on the correlation between pre pregnancy BMI and weight changes during pregnancy, China Maternal and Child Health Care, 2017, 32 (15): 3360-3364.

[4] Zhang Rui, Zhang Hui, Li Yan, et al. External quality assessment of fetuses with trisomy 21, trisomy 18, and trisomy 13 detected by large-scale parallel sequencing in clinical laboratories. Journal of Molecular Diagnosis, March 2016, 18 (2): 244-252.

[5] Chiu R W, Akolekar R, Zheng Y W, et al. Non-invasive prenatal assessment of trisomy 21 by multiplexed maternal plasma DNA sequencing: large scale validity study. BMJ, 2011, 342: c7401.

[6] Norton M E, Brar H, Weiss J, et al. Non-Invasive Prenatal Testing of Chromosomal Aneuploidy: Clinical Experience and Future Directions. Prenatal diagnosis, 2015, 35(1): 1-18.

[7] Dan S, Wang W, Ren J, et al. Clinical application of maternal plasma DNA analysis for noninvasive prenatal diagnosis of trisomy 21: systematic review and meta-analysis. BMJ open, 2013, 3(1): e002468.

[8] Chiu R W, Akolekar R, Zheng Y W. Non-invasive prenatal assessment of trisomy 21 by multiplexed maternal plasma DNA sequencing: updated large-scale validity study. BMJ, 2023, 381: e071234.