Academic Journal of Mathematical Sciences, 2025, 6(2); doi: 10.25236/AJMS.2025.060216.
Panlin Li, Ning Zhang
College of Mathematics and Physics, Xinjiang Agricultural University, Urumqi, China
To address the challenge of determining chromosomal abnormalities in female fetuses, this study developed a scientific diagnostic model by comprehensively analyzing multiple factors—including Z-scores, GC content, read counts, relative proportions, and BMI—of the X chromosome and chromosomes 21, 18, and 13 in both pregnant women and fetuses, despite the absence of the Y chromosome as a reference. To resolve sample category imbalance, SMOTE oversampling technology was employed to expand the minority category to match the majority category in scale. Regarding modeling approaches, three distinct model systems were constructed: logistic regression, Probit regression, and adaptive norm robust probabilistic regression. Experimental results indicate that logistic regression performed best in detecting T13 abnormalities (76.0% accuracy, AUC=0.821); Probit regression demonstrated superiority in marginal effect interpretation (76.0% accuracy, AUC=0.739); while the adaptive norm robust model achieved a high accuracy of 87.5% in T13 anomaly detection. Feature importance analysis indicated that GC content played a dominant role in identifying all anomaly types. The findings validate the effectiveness and interpretability of the developed models, providing novel methodological support for non-invasive prenatal testing.
Female Fetal Chromosomal Abnormalities; SMOTE Oversampling; Logistic Regression; Probit Regression; Robust Probability Regression; GC Content
Panlin Li, Ning Zhang. Study on the Determination of Fetal Chromosomal Abnormalities Based on Multi-Model Regression. Academic Journal of Mathematical Sciences (2025), Vol. 6, Issue 2: 121-129. https://doi.org/10.25236/AJMS.2025.060216.
[1] Zhu Q, Wang J, Xu X, Zhou S, Liao Z, Zhang J, Kong L, Liang B, Cheng X. KF-NIPT: K-mer and fetal fraction-based estimation of chromosomal anomaly from NIPT data[J]. BMC Bioinformatics, 2025, 26(1): 15.
[2] Staniczek J , Manasar-Dyrbu M , Sodowska P ,et al.Fetal karyotyping in adolescent pregnancies: a population-based cohort study on outcomes of invasive prenatal testing[J].Frontiers in Genetics, 2025, 16(16):1581249.DOI:10.3389/fgene.2025.1581249.
[3] Zhang Y, Xu H, Zhang W, Liu K, et al. Non-invasive prenatal testing for the detection of trisomy 13, 18, and 21 and sex chromosome aneuploidies in 68,763 cases[J]. Front Genet, 2022, 13: 857421.
[4] Wei R , Hu H , Wang S ,et al.Analysis of chromosomal aberrations in early pregnancy loss using high-throughput ligation-dependent probe amplification and single tandem repeats[J].Molecular Cytogenetics, 2025, 18.DOI:10.1186/s13039-025-00724-5.
[5] Ye Y , Ma J , Zhan Q ,et al.Controlled Ovarian Stimulation Contributes to the Incidence of de Novo Chromosomal Abnormalities in Cleavage-Stage Embryos[J].Archives of Medical Research, 2025, 57(3).DOI:10.1016/j.arcmed.2025.103318.
[6] Yanchun Z, Hongyan X, Wen Z, Kaibo L, et al. A machine learning technology to improve the risk of non-invasive prenatal tests[J]. BMC Med Genomics, 2023, 16(1): 54.
[7] Nie J, Wang Y, Li F, Chen X, et al. Development and Validation of a Deep Learning Model to Screen for Trisomy 21 During the First Trimester From Nuchal Ultrasonographic Images[J]. JAMA Netw Open, 2022, 5(12): e2248487.
[8] Lee S, Park J, Kim H, et al. Accuracy of cell-free fetal DNA in detecting chromosomal anomalies in women experiencing miscarriage: systematic review and meta-analysis[J]. Ultrasound Obstet Gynecol, 2024, 63(5): 678-687.
[9] Al-Ghaili A M, et al. Artificial intelligence for prenatal chromosome analysis: a systematic review[J]. Comput Biol Med, 2023, 158: 106860.
[10] Chen Y, Wang H, Li J, et al. Application of intelligent algorithms in Down syndrome screening during second trimester pregnancy[J]. BMC Pregnancy Childbirth, 2023, 23(1): 115.
[11] Smith R, Johnson T, Liu H, et al. Prenatal Screening for Chromosomal Defects[v1][J]. Preprints, 2025: 2025030671.
[12] Zhao X, Chen L, et al. Novel method of real-time PCR-based screening for common fetal trisomies[J]. BMC Med Genomics, 2021, 14(1): 36.
[13] Feng C, Liu W. Scalable Bayesian p-generalized probit and logistic regression[J]. Adv Data Anal Classif, 2024, 18: 317-334.
[14] Hossain M, Rahman M, et al. Robust adaptive Lasso in high-dimensional logistic regression[J]. Stat Methods Med Res, 2021, 30(9): 2019-2034.
[15] Li X, Zhou Y, Zhang H, et al. A novel framework for abnormal risk classification over fetal nuchal translucency using adaptive stochastic gradient descent algorithm[J]. Diagnostics, 2022, 12(3): 682.