Limits...
Statistical Approach to Decreasing the Error Rate of Noninvasive Prenatal Aneuploid Detection caused by Maternal Copy Number Variation.

Zhang H, Zhao YY, Song J, Zhu QY, Yang H, Zheng ML, Xuan ZL, Wei Y, Chen Y, Yuan PB, Yu Y, Li DW, Liang JB, Fan L, Chen CJ, Qiao J - Sci Rep (2015)

Bottom Line: We also introduced a method based on the discrepancy between feto-placental concentrations to help reduce the false-positive ratio.All 106 fetuses with T21, 20 with T18, and three with T13 were tested using our method, with sensitivity of 100% and specificity of 99.97%.In the results, two cases with maternal duplications in chromosome 21, which were falsely predicted as T21 by the previous NIPT method, were correctly classified as normal by our algorithm, which demonstrated the effectiveness of our approach.

View Article: PubMed Central - PubMed

Affiliation: Annoroad Gene Technology Co., Ltd, Beijing, China.

ABSTRACT
Analyses of cell-free fetal DNA (cff-DNA) from maternal plasma using massively parallel sequencing enable the noninvasive detection of feto-placental chromosome aneuploidy; this technique has been widely used in clinics worldwide. Noninvasive prenatal tests (NIPT) based on cff-DNA have achieved very high accuracy; however, they suffer from maternal copy-number variations (CNV) that may cause false positives and false negatives. In this study, we developed an algorithm to exclude the effect of maternal CNV and refined the Z-score that is used to determine fetal aneuploidy. The simulation results showed that the algorithm is robust against variations of fetal concentration and maternal CNV size. We also introduced a method based on the discrepancy between feto-placental concentrations to help reduce the false-positive ratio. A total of 6615 pregnant women were enrolled in a prospective study to validate the accuracy of our method. All 106 fetuses with T21, 20 with T18, and three with T13 were tested using our method, with sensitivity of 100% and specificity of 99.97%. In the results, two cases with maternal duplications in chromosome 21, which were falsely predicted as T21 by the previous NIPT method, were correctly classified as normal by our algorithm, which demonstrated the effectiveness of our approach.

No MeSH data available.


Simulation results for maternal duplication (A) and maternal deletion (B) in chromosomes 13, 18 and 21. olid lines represent raw Z-scores of the simulation result and dashed lines indicate Z-scores after our maternal CNV (MAT-CNV) adjustment. The black, red and green lines represent fetal concentrations of 5%, 10% and 15% respectively. The x-axis indicates the size of maternal CNV, while the y-axis shows the Z-score.
© Copyright Policy - open-access
Related In: Results  -  Collection

License
getmorefigures.php?uid=PMC4632076&req=5

f1: Simulation results for maternal duplication (A) and maternal deletion (B) in chromosomes 13, 18 and 21. olid lines represent raw Z-scores of the simulation result and dashed lines indicate Z-scores after our maternal CNV (MAT-CNV) adjustment. The black, red and green lines represent fetal concentrations of 5%, 10% and 15% respectively. The x-axis indicates the size of maternal CNV, while the y-axis shows the Z-score.

Mentions: We subsequently investigated the effect of maternal CNVs on the final Z-score calculation. As demonstrated in Fig. 1, the raw Z-scores, before revision by our MAT-CNV approach, increased or decreased in proportion with the size of the maternal duplication or deletion respectively, although the degree of correlation varied in terms of fetal concentrations and different chromosomes. Higher fetal DNA fraction and longer valid chromosome length reduced the influence of maternal CNV on the raw Z-score. Apparently, a maternal duplication of more than 1.5 Mb on chromosome 21, whatever the fetal concentration, resulted in a Z-score larger than 3 in a euploid fetus, leading to a false-positive NIPT result. The thresholds of maternal CNV size that will cause a false-positive NIPT result in chromosomes 18 and 13 are 2.2 Mb and 3.2 Mb, respectively. Remarkably, after applying the MAT-CNV approach, the Z-scores remained around 0 under all circumstances, implying that our approach was successful in reducing discrepant NIPT results caused by maternal CNVs.


Statistical Approach to Decreasing the Error Rate of Noninvasive Prenatal Aneuploid Detection caused by Maternal Copy Number Variation.

Zhang H, Zhao YY, Song J, Zhu QY, Yang H, Zheng ML, Xuan ZL, Wei Y, Chen Y, Yuan PB, Yu Y, Li DW, Liang JB, Fan L, Chen CJ, Qiao J - Sci Rep (2015)

Simulation results for maternal duplication (A) and maternal deletion (B) in chromosomes 13, 18 and 21. olid lines represent raw Z-scores of the simulation result and dashed lines indicate Z-scores after our maternal CNV (MAT-CNV) adjustment. The black, red and green lines represent fetal concentrations of 5%, 10% and 15% respectively. The x-axis indicates the size of maternal CNV, while the y-axis shows the Z-score.
© Copyright Policy - open-access
Related In: Results  -  Collection

License
Show All Figures
getmorefigures.php?uid=PMC4632076&req=5

f1: Simulation results for maternal duplication (A) and maternal deletion (B) in chromosomes 13, 18 and 21. olid lines represent raw Z-scores of the simulation result and dashed lines indicate Z-scores after our maternal CNV (MAT-CNV) adjustment. The black, red and green lines represent fetal concentrations of 5%, 10% and 15% respectively. The x-axis indicates the size of maternal CNV, while the y-axis shows the Z-score.
Mentions: We subsequently investigated the effect of maternal CNVs on the final Z-score calculation. As demonstrated in Fig. 1, the raw Z-scores, before revision by our MAT-CNV approach, increased or decreased in proportion with the size of the maternal duplication or deletion respectively, although the degree of correlation varied in terms of fetal concentrations and different chromosomes. Higher fetal DNA fraction and longer valid chromosome length reduced the influence of maternal CNV on the raw Z-score. Apparently, a maternal duplication of more than 1.5 Mb on chromosome 21, whatever the fetal concentration, resulted in a Z-score larger than 3 in a euploid fetus, leading to a false-positive NIPT result. The thresholds of maternal CNV size that will cause a false-positive NIPT result in chromosomes 18 and 13 are 2.2 Mb and 3.2 Mb, respectively. Remarkably, after applying the MAT-CNV approach, the Z-scores remained around 0 under all circumstances, implying that our approach was successful in reducing discrepant NIPT results caused by maternal CNVs.

Bottom Line: We also introduced a method based on the discrepancy between feto-placental concentrations to help reduce the false-positive ratio.All 106 fetuses with T21, 20 with T18, and three with T13 were tested using our method, with sensitivity of 100% and specificity of 99.97%.In the results, two cases with maternal duplications in chromosome 21, which were falsely predicted as T21 by the previous NIPT method, were correctly classified as normal by our algorithm, which demonstrated the effectiveness of our approach.

View Article: PubMed Central - PubMed

Affiliation: Annoroad Gene Technology Co., Ltd, Beijing, China.

ABSTRACT
Analyses of cell-free fetal DNA (cff-DNA) from maternal plasma using massively parallel sequencing enable the noninvasive detection of feto-placental chromosome aneuploidy; this technique has been widely used in clinics worldwide. Noninvasive prenatal tests (NIPT) based on cff-DNA have achieved very high accuracy; however, they suffer from maternal copy-number variations (CNV) that may cause false positives and false negatives. In this study, we developed an algorithm to exclude the effect of maternal CNV and refined the Z-score that is used to determine fetal aneuploidy. The simulation results showed that the algorithm is robust against variations of fetal concentration and maternal CNV size. We also introduced a method based on the discrepancy between feto-placental concentrations to help reduce the false-positive ratio. A total of 6615 pregnant women were enrolled in a prospective study to validate the accuracy of our method. All 106 fetuses with T21, 20 with T18, and three with T13 were tested using our method, with sensitivity of 100% and specificity of 99.97%. In the results, two cases with maternal duplications in chromosome 21, which were falsely predicted as T21 by the previous NIPT method, were correctly classified as normal by our algorithm, which demonstrated the effectiveness of our approach.

No MeSH data available.