Noninvasive prenatal testing for fetal subchromosomal copy number variations and chromosomal aneuploidy by low‐pass whole‐genome sequencing

Dongyi Yu, Kai Zhang, Meiyan Han, Wei Pan, Ying Chen, Yunfeng Wang, Jiao Hongyan, Ling Duan, Qiying Zhu, Xiaojie Song, Yan Hong, Juan Wang, Hui Feng, Linzhou Huang, Chongjian Chen, Yang DuMolecular 2019 · DOI

TL;DR

This paper introduces NIPSCCD, a novel low-pass whole-genome sequencing (WGS) method for noninvasively detecting fetal subchromosomal copy number variations (CNVs) and chromosomal aneuploidies from maternal plasma. The method demonstrates high sensitivity and specificity, particularly for larger CNVs, and can be integrated into routine NIPT workflows, offering a more comprehensive prenatal screening.

What Problem Does This Paper Address?

Noninvasive prenatal testing (NIPT) has revolutionized prenatal screening by using cell-free fetal DNA (cffDNA) from maternal blood to detect common chromosomal aneuploidies like Down syndrome. However, extending NIPT to detect subchromosomal CNVs (segmental deletions and duplications) has been challenging. These CNVs are associated with a range of severe genetic disorders and occur in a significant percentage of fetuses with structural anomalies. While invasive methods like amniocentesis can detect these CNVs, they carry risks. Previous attempts at noninvasive CNV detection using NIPT have faced limitations in sensitivity and specificity, often requiring deep sequencing that is not cost-effective for widespread clinical use. This paper aims to bridge this gap by developing and validating a low-pass WGS approach for accurate and efficient detection of both chromosomal aneuploidies and subchromosomal CNVs.

Methods

The core of this study is the development of the Noninvasively Prenatal Subchromosomal Copy number variation Detection (NIPSCCD) method. This approach utilizes low-pass whole-genome sequencing of maternal plasma. The workflow involves:

  1. Genome Segmentation: Dividing the genome into 100-Kb windows.
  2. Read Counting and Correction: Collecting unique reads per window, correcting for GC content bias using a LOWESS model, and capping values to reduce noise.
  3. Z-score Transformation: Calculating raw Z-scores for each window and then transforming them against a reference set of low-risk samples to obtain normalized Z-scores.
  4. CNV Calling: Merging consecutive windows and applying the Circular Binary Segmentation (CBS) algorithm to identify segments with absolute Z-scores above a threshold (1.28 for CNVs, 3 for whole chromosome aneuploidies).
  5. Fetal Fraction Estimation: Estimating fetal fraction using reads mapped to the Y chromosome and an alternative depth-based model for CNV regions. Discrepancies between these estimations were used to filter false positives.
  6. Karyotype Calling: Calling CNVs and fetal karyotypes based on pre-defined thresholds.

The key innovation lies in using low-pass WGS and developing a robust bioinformatic pipeline (NIPSCCD) that can achieve high performance without the need for deep sequencing, making it more clinically and economically viable.

Datasets

The study evaluated the NIPSCCD method using samples from 20,003 pregnant women. These samples were collected between July 2015 and June 2016. For validation, confirmatory results from amniocentesis followed by whole-genome sequencing or chromosomal microarray analysis were used as the gold standard for CNVs. For chromosomal aneuploidies, karyotyping results were used. The paper does not explicitly mention the creation of a new benchmark dataset but rather uses existing clinical samples for validation against established diagnostic methods.

Key Insights and Ideas

  • Unified Screening Power: NIPSCCD successfully integrates the detection of both common chromosomal aneuploidies and subchromosomal CNVs into a single, low-pass WGS assay. This offers a more comprehensive prenatal screening than current NIPT methods.
  • Size Matters for CNV Detection: The sensitivity of NIPSCCD for detecting CNVs is strongly influenced by their size. Larger CNVs (>10 Mb) are detected with significantly higher accuracy than smaller ones (<5 Mb). This is a critical consideration for clinical interpretation.
  • Low-Pass WGS is Sufficient: The study demonstrates that even low-pass WGS, when coupled with a sophisticated bioinformatic pipeline like NIPSCCD, can achieve clinically relevant detection rates for CNVs, challenging the notion that deep sequencing is always necessary.
  • Potential for Routine Integration: The high sensitivity and specificity for both aneuploidies and CNVs suggest that NIPSCCD is a strong candidate for integration into routine NIPT programs, enhancing the diagnostic yield without substantially increasing costs or complexity.

Key Results

  • CNV Detection: NIPSCCD identified 36 CNVs that were confirmed by amniocentesis testing.
  • Sensitivity for CNVs >10 Mb: 91.67%
    • Sensitivity for CNVs 5 Mb–10 Mb: 100.00%
    • Sensitivity for CNVs <5 Mb: 68.42%
    • Overall sensitivity for CNVs: 80.56%
    • Overall PPV for CNVs: 80.56%
  • Aneuploidy Detection: NIPSCCD identified 124 positive trisomy cases (21, 18, 13), with 103 confirmed by karyotyping.
  • Overall sensitivity for trisomy 21/18/13: 100.00%
    • Overall specificity for trisomy 21/18/13: 99.89%

Limitations and Future Work

The authors acknowledge several limitations:

  • CNV Size Dependency: The lower sensitivity for smaller CNVs (<5 Mb) remains a challenge. Six out of seven false-negative CNVs were less than 2.5 Mb, highlighting the need for further optimization to improve detection of smaller variants.
  • Validation of Unknown Significance CNVs: A significant number of women with NIPSCCD-identified CNVs of unknown clinical significance declined invasive procedures for validation. This limits the precise assessment of false positive rates for these cases.
  • Confirmatory Testing Method: Most positive CNVs were validated by low-pass WGS of amniocytes, not chromosomal microarray analysis. While WGS is powerful, potential differences in sensitivity between these methods could influence the accuracy of validation.
  • Maternal CNV Differentiation: The study notes the inherent difficulty in differentiating maternal CNVs from fetal CNVs when the fetus is female, which can contribute to false positives.

Future work should focus on optimizing the pipeline for smaller CNVs, developing more robust methods for differentiating maternal and fetal CNVs, and potentially incorporating maternal background testing to improve diagnostic accuracy and provide comprehensive genetic counseling.

Summary updated

Abstract

NIPSCCD showed a good performance in detecting fetal subchromosomal CNVs, especially for CNVs >10 Mb, and can be incorporated into the routine NIPT chromosomal aneuploidies screening with high sensitivity and specificity.
BibTeX
@inproceedings{yu2019noninvasive,
  author = {Dongyi Yu and Kai Zhang and Meiyan Han and Wei Pan and Ying Chen and Yunfeng Wang and Jiao Hongyan and Ling Duan and Qiying Zhu and Xiaojie Song and Yan Hong and Chen Chen and Juan Wang and Hui Feng and Linzhou Huang and Chongjian Chen and Yang Du},
  title = {{Noninvasive prenatal testing for fetal subchromosomal copy number variations and chromosomal aneuploidy by low‐pass whole‐genome sequencing}},
  booktitle = {Molecular},
  year = {2019},
  doi = {10.1002/mgg3.674},
}
Figure