Journal Article

FZJ-2018-00010

http://join2-wiki.gsi.de/foswiki/pub/Main/Artwork/join2_logo100x88.png Analysis of the joint effect of SNPs to identify independent loci and allelic heterogeneity in schizophrenia GWAS data

Polushina, T. ; Giddaluru, S. ; Bettella, F. ; Espeseth, T. ; Lundervold, A. J. ; Djurovic, S. ; Cichon, S.FZJ* ; Hoffmann, P. ; Nöthen, M. M. ; Steen, V. M. ; Andreassen, O. A. ; Le Hellard, S. (Corresponding author)

2017
Nature Publishing Group London

This record in other databases:      

Please use a persistent id in citations: http://hdl.handle.net/2128/16338  doi:10.1038/s41398-017-0033-2

Abstract: We have tested published methods for capturing allelic heterogeneity and identifying loci of joint effects to uncover more of the “hidden heritability” of schizophrenia (SCZ). We used two tools, cojo-GCTA and multi-SNP, to analyze meta-statistics from the latest genome-wide association study (GWAS) on SCZ by the Psychiatric Genomics Consortium (PGC). Stepwise regression on markers with p values <10−7 in cojo-GCTA identified 96 independent signals. Eighty-five passed the genome-wide significance threshold. Cross-validation of cojo-GCTA by CLUMP was 76%, i.e., 26 of the loci identified by the PGC using CLUMP were found to be dependent on another locus by cojo-GCTA. The overlap between cojo-GCTA and multi-SNP was better (up to 92%). Three markers reached genome-wide significance (5 × 10−8) in a joint effect model. In addition, two loci showed possible allelic heterogeneity within 1-Mb genomic regions, while CLUMP analysis had identified 16 such regions. Cojo-GCTA identified fewer independent loci than CLUMP and seems to be more conservative, probably because it accounts for long-range LD and interaction effects between markers. These findings also explain why fewer loci with possible allelic heterogeneity remained significant after cojo-GCTA analysis. With multi-SNP, 86 markers were selected at the threshold 10−7. Multi-SNP identifies fewer independent signals, due to splitting of the data and use of smaller samples. We recommend that cojo-GCTA and multi-SNP are used for post-GWAS analysis of all traits to call independent loci. We conclude that only a few loci in SCZ show joint effects or allelic heterogeneity, but this could be due to lack of power for that data set.

---

Contributing Institute(s):

1. Strukturelle und funktionelle Organisation des Gehirns (INM-1)

Research Program(s):

1. 571 - Connectivity and Activity (POF3-571) (POF3-571)

Appears in the scientific report 2017

---

Database coverage:
; ; ; ; Current Contents - Clinical Medicine ; DOAJ Seal ; IF >= 5 ; JCR ; NCBI Molecular Biology Database ; SCOPUS ; Science Citation Index Expanded ; Thomson Reuters Master Journal List ; Web of Science Core Collection

---