Abstract
Motivation: Copy number variants (CNVs) have been implicated in a variety of neurodevelopmental disorders, including autism spectrum disorders, intellectual disability and schizophrenia. Recent advances in high-throughput genomic technologies have enabled rapid discovery of many genetic variants including CNVs. As a result, there is increasing interest in studying the role of CNVs in the etiology of many complex diseases. Despite the availability of an unprecedented wealth of CNV data, methods for testing association between CNVs and disease-related traits are still underdeveloped due to the low prevalence and complicated multi-scale features of CNVs. Results: We propose a novel CNV kernel association test (CKAT) in this paper. To address the low prevalence, CNVs are first grouped into CNV regions (CNVR). Then, taking into account the multiscale features of CNVs, we first design a single-CNV kernel which summarizes the similarity between two CNVs, and next aggregate the single-CNV kernel to a CNVR kernel which summarizes the similarity between two CNVRs. Finally, association between CNVR and disease-related traits is assessed by comparing the kernel-based similarity with the similarity in the trait using a score test for variance components in a random effect model.We illustrate the proposed CKAT using simulations and show that CKAT is more powerful than existing methods, while always being able to control the type I error. We also apply CKAT to a real dataset examining the association between CNV and autism spectrum disorders, which demonstrates the potential usefulness of the proposed method.
| Original language | English (US) |
|---|---|
| Pages (from-to) | 3603-3610 |
| Number of pages | 8 |
| Journal | Bioinformatics |
| Volume | 32 |
| Issue number | 23 |
| DOIs | |
| State | Published - Jan 1 2016 |
Fingerprint
All Science Journal Classification (ASJC) codes
- Statistics and Probability
- Biochemistry
- Molecular Biology
- Computer Science Applications
- Computational Theory and Mathematics
- Computational Mathematics
Cite this
}
A novel copy number variants kernel association test with application to autism spectrum disorders studies. / Zhan, Xiang; Girirajan, Santhosh; Zhao, Ni; Wu, Michael C.; Ghosh, Debashis.
In: Bioinformatics, Vol. 32, No. 23, 01.01.2016, p. 3603-3610.Research output: Contribution to journal › Article
TY - JOUR
T1 - A novel copy number variants kernel association test with application to autism spectrum disorders studies
AU - Zhan, Xiang
AU - Girirajan, Santhosh
AU - Zhao, Ni
AU - Wu, Michael C.
AU - Ghosh, Debashis
PY - 2016/1/1
Y1 - 2016/1/1
N2 - Motivation: Copy number variants (CNVs) have been implicated in a variety of neurodevelopmental disorders, including autism spectrum disorders, intellectual disability and schizophrenia. Recent advances in high-throughput genomic technologies have enabled rapid discovery of many genetic variants including CNVs. As a result, there is increasing interest in studying the role of CNVs in the etiology of many complex diseases. Despite the availability of an unprecedented wealth of CNV data, methods for testing association between CNVs and disease-related traits are still underdeveloped due to the low prevalence and complicated multi-scale features of CNVs. Results: We propose a novel CNV kernel association test (CKAT) in this paper. To address the low prevalence, CNVs are first grouped into CNV regions (CNVR). Then, taking into account the multiscale features of CNVs, we first design a single-CNV kernel which summarizes the similarity between two CNVs, and next aggregate the single-CNV kernel to a CNVR kernel which summarizes the similarity between two CNVRs. Finally, association between CNVR and disease-related traits is assessed by comparing the kernel-based similarity with the similarity in the trait using a score test for variance components in a random effect model.We illustrate the proposed CKAT using simulations and show that CKAT is more powerful than existing methods, while always being able to control the type I error. We also apply CKAT to a real dataset examining the association between CNV and autism spectrum disorders, which demonstrates the potential usefulness of the proposed method.
AB - Motivation: Copy number variants (CNVs) have been implicated in a variety of neurodevelopmental disorders, including autism spectrum disorders, intellectual disability and schizophrenia. Recent advances in high-throughput genomic technologies have enabled rapid discovery of many genetic variants including CNVs. As a result, there is increasing interest in studying the role of CNVs in the etiology of many complex diseases. Despite the availability of an unprecedented wealth of CNV data, methods for testing association between CNVs and disease-related traits are still underdeveloped due to the low prevalence and complicated multi-scale features of CNVs. Results: We propose a novel CNV kernel association test (CKAT) in this paper. To address the low prevalence, CNVs are first grouped into CNV regions (CNVR). Then, taking into account the multiscale features of CNVs, we first design a single-CNV kernel which summarizes the similarity between two CNVs, and next aggregate the single-CNV kernel to a CNVR kernel which summarizes the similarity between two CNVRs. Finally, association between CNVR and disease-related traits is assessed by comparing the kernel-based similarity with the similarity in the trait using a score test for variance components in a random effect model.We illustrate the proposed CKAT using simulations and show that CKAT is more powerful than existing methods, while always being able to control the type I error. We also apply CKAT to a real dataset examining the association between CNV and autism spectrum disorders, which demonstrates the potential usefulness of the proposed method.
UR - http://www.scopus.com/inward/record.url?scp=85016169469&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85016169469&partnerID=8YFLogxK
U2 - 10.1093/bioinformatics/btw500
DO - 10.1093/bioinformatics/btw500
M3 - Article
VL - 32
SP - 3603
EP - 3610
JO - Bioinformatics
JF - Bioinformatics
SN - 1367-4803
IS - 23
ER -