A Method to Create NIPT Samples with Turner Disorder to Evaluate NIPT Algorithms

Abstract

Noninvasive prenatal test (NIPT) is a widely used screening method to detect numerical aberrations on autosomal chromosomes with high sensitivity and specificity. However, predicting abnormalities on sex chromosomes is much more challenging due to complicated genomic characteristics of sex chromosomes. Turner disorder (XO) is one of the most common disorders due to the missing of one X chromosome in females. A number of large-scale retrospective studies have showed that the positive detection rate of Turner disorder is considerably high and the false negative rate has not been well evaluated due to the lack of available positive Turner samples. To solve the problem, we present a novel method to create positive Turner samples from negative samples that can be easily obtained from NIPT testing centers. We applied the method to create 600 positive Turner samples and examined and the performance of WisecondorX, CNVkit, and VINIPT algorithms on the samples. Experiments show that the sensitivity of WisecondorX, VINIPT, and CNVkit in detecting positive Turner samples are 100%, 100%, and 99.5%, respectively. We also evaluated the performance of the algorithms on 500 negative XO samples. The VINIPT and CNVkit algorithms have very high specificity in identifying negative XO samples (i.e., 99.8% for VINIPT and 99.6% for CNVkit), while WisecondorX has a lower specificity of 96.8%. The study opens an easy way for researchers to assess the performance of NIPT algorithms on screening the Turner disorder.
 Keywords: NIPT, cfDNA analysis, Turner disorder, XO aberration, WisecondorX, CNVkit.