Million Veterans Program
The Million Veterans Program [
51] (MVP) biobank is one of the world’s leading repositories of genetic and phenotypic information, and is an unprecedented resource for the study of PTSD. A conference abstract on MVP GWAS of PTSD re-experience symptoms has been published [
12••] and additional results about other PTSD phenotypes measured within MVP will be available in future publications. Regarding PTSD re-experiencing symptoms, MVP researchers examined a sample of 146,660 European-ancestry participants and 19,983 African-ancestry participants. This dataset afforded the discovery of eight loci at the level of genome-wide significance (i.e.,
p < 5 × 10
−8). These loci include a chromosome 3 locus with top variant rs2777888 (
p = 2.1 × 10
−11). This variant is located in an intron of the
CAMKV gene (CaM kinase like vesicle associated), which is highly expressed in the brain. An extended locus on chromosome 17 was also identified, with lead SNP rs2532252 (
p = 4.5 × 10
−10) closest to the
KANSL1 gene (KAT8 regulatory NSL complex subunit 1). This locus also encompasses the
CRHR1 gene (corticotropin releasing hormone receptor 1), a previous candidate gene for PTSD. This means that CRHR1 may be associated with PTSD, but further research is needed to assess this possibility given the large number of genes and regulatory regions in this broad locus (see notes about fine-mapping below). A third locus on chromosome 18 is located in a locus previously associated with schizophrenia, in the
TCF4 gene (transcription factor 4; top PTSD SNP rs2123392,
p = 5.4 × 10
−11). The discovery of these loci [
12••] is a tremendous step forward for PTSD genetics. The forthcoming full manuscript, as well as further genetic studies of PTSD phenotypes from MVP, will provide motivating results for the field. The major limitation of the MVP studies is the exclusive focus on military samples. The Psychiatric Genomics Consortium studies (below) include both military and civilian samples, from a wide variety of contexts relevant to PTSD.
PTSD Group of the Psychiatric Genomics Consortium (PGC-PTSD) and the UK Biobank
As has been the case for much of complex trait genetics research, the formation of international consortia focused on the genetics of PTSD has been a critical step in for discovery because far larger sample sizes can be achieved through sharing of data. In psychiatry, the largest genetics consortium is the Psychiatric Genomics Consortium, PGC [
41]. The PGC made possible the identification of over 100 risk loci for schizophrenia, as reported in 2014 [
37], and more loci have consistently been identified as data aggregation within the PGC schizophrenia group has continued. The PGC-PTSD group [
21,
22] has been employing the same strategy used by other successful PGC groups, and the first empirical paper for the group had a sample size of 20,070 [
11••], from the combined analysis of 11 previous GWAS of PTSD [
52‐
56]. This study was notable because it was the first to be adequately powered to estimate
h2SNP for PTSD. Intriguingly, the
h2SNP estimate for females (29%) was higher than the
h2SNP estimate for males (7%), consistent with twin study PTSD heritability estimates (i.e.,
h2twin_FEMALE estimates are higher than
h2twin_MALE estimates), as described above. Analyses from the PGC-PTSD group also revealed shared genetic effects between PTSD and schizophrenia, bipolar disorder, and depression [
11••,
13••].
The second wave of data analysis from the PGC-PTSD group (abstract currently available [
13••], and manuscript forthcoming) replicated and extended the findings from the first PGC-PTSD paper [
11••] and also identified potential specific risk loci and genes for PTSD. At the time of writing of this review, the following information about top loci is available from the published abstract [
13••]. Stratified analyses revealed two loci (on chromosomes 6 and 13) that exceeded genome-wide significance in the European ancestry analyses (6q25,
p = 3.1 × 10
−9 and 13q32,
p = 2.7 × 10
−8). In the African ancestry analyses, a separate locus on chromosome 13 exceeded genome-wide significance (13q.21,
p = 3.8 × 10
−8). Further, polygenic analyses make it clear that the identification of many more loci will occur once adequate power is achieved. Thus, sample collection is continuing within the PGC-PTSD group. Like those loci identified by in the Million Veterans Program, the next step for the PGC-PTSD loci is “fine-mapping.” Fine-mapping [
57•] refers to various analytical and biological procedures used to refine the signal within a particular locus, ideally to the resolution of individual variants causally associated with disease.
In closing this section about GWAS results, we note that the methodological advantages of GWAS (described above) do not imply that GWAS results should be accepted indiscriminately. Rather, consumers of the GWAS literature should be aware of certain guidelines in the evaluation of GWAS results. Above all, sample size has proven to be the best indicator of how many loci will be discovered and how robust findings will prove to be upon investigation in novel samples. Thus, for a given phenotype such as PTSD, a good rule of thumb is that the largest GWAS (i.e., largest N) will likely provide the best information about molecular genetic influences on PTSD. Another indicator of power in GWAS is the presence of a significant SNP-heritability estimate (h2SNP).
In general, smaller GWASs can be used to conduct polygenic analyses of heritability and genetic correlations, than are necessary for the identification of individual risk loci [
23,
58]. Sample sizes of many tens of thousands of participants have been necessary for risk locus discovery [
26••,
37], whereas methods like GCTA [
23] can be used to estimate
h2SNP using just thousands of samples. For these reasons, we chose to focus on the MVP and PGC-PTSD GWAS results instead of smaller GWAS studies, which were not even adequately powered for heritability analyses (and by extension, it is less likely that they were adequately powered to detect individual loci). At the same time, it is important to recognize that the progress made by GWAS consortia in recent years would not have been possible without the considerable efforts involved in the conduct of each individual GWAS study. The individual GWAS that made recent consortium findings possible are provided in the reference list [
52‐
56,
59‐
66] and described in detail in the PGC papers.