Title |
An adaptive detection method for fetal chromosomal aneuploidy using cell-free DNA from 447 Korean women
|
---|---|
Published in |
BMC Medical Genomics, October 2016
|
DOI | 10.1186/s12920-016-0222-5 |
Pubmed ID | |
Authors |
Sunshin Kim, Sunshin Kim, HeeJung Jung, Sung Hee Han, SeungJae Lee, JeongSub Kwon, Min Gyun Kim, Hyungsik Chu, Kyudong Han, Hwanjong Kwak, Sunghoon Park, Hee Jae Joo, Minae An, Jungsu Ha, Kyusang Lee, Byung Chul Kim, Hailing Zheng, Xinqiang Zhu, Hongliang Chen, Jong Bhak |
Abstract |
Noninvasive prenatal testing (NIPT) using massively parallel sequencing of cell-free DNA (cfDNA) is increasingly being used to predict fetal chromosomal abnormalities. However, concerns over erroneous predictions which occur while performing NIPT still exist in pregnant women at high risk for fetal aneuploidy. We performed the largest-scale clinical NIPT study in Korea to date to assess the risk of false negatives and false positives using next-generation sequencing. A total of 447 pregnant women at high risk for fetal aneuploidy were enrolled at 12 hospitals in Korea. They underwent definitive diagnoses by full karyotyping by blind analysis and received aneuploidy screening at 11-22 weeks of gestation. Three steps were employed for cfDNA analyses. First, cfDNA was sequenced. Second, the effect of GC bias was corrected using normalization of samples as well as LOESS and linear regressions. Finally, statistical analysis was performed after selecting a set of reference samples optimally adapted to a test sample from the whole reference samples. We evaluated our approach by performing cfDNA testing to assess the risk of trisomies 13, 18, and 21 using the sets of extracted reference samples. The adaptive selection algorithm presented here was used to choose a more optimized reference sample, which was evaluated by the coefficient of variation (CV), demonstrated a lower CV and higher sensitivity than standard approaches. Our adaptive approach also showed that fetal aneuploidies could be detected correctly by clearly splitting the z scores obtained for positive and negative samples. We show that our adaptive reference selection algorithm for optimizing trisomy detection showed improved reliability and will further support practitioners in reducing both false negative and positive results. |
X Demographics
Geographical breakdown
Country | Count | As % |
---|---|---|
United Kingdom | 1 | 100% |
Demographic breakdown
Type | Count | As % |
---|---|---|
Scientists | 1 | 100% |
Mendeley readers
Geographical breakdown
Country | Count | As % |
---|---|---|
Unknown | 28 | 100% |
Demographic breakdown
Readers by professional status | Count | As % |
---|---|---|
Researcher | 6 | 21% |
Student > Ph. D. Student | 4 | 14% |
Other | 3 | 11% |
Student > Bachelor | 2 | 7% |
Professor > Associate Professor | 2 | 7% |
Other | 4 | 14% |
Unknown | 7 | 25% |
Readers by discipline | Count | As % |
---|---|---|
Medicine and Dentistry | 7 | 25% |
Biochemistry, Genetics and Molecular Biology | 6 | 21% |
Computer Science | 3 | 11% |
Agricultural and Biological Sciences | 2 | 7% |
Pharmacology, Toxicology and Pharmaceutical Science | 1 | 4% |
Other | 1 | 4% |
Unknown | 8 | 29% |