A paper1 published this week in Nature marks the culmination of a long debate about the genetic basis of a disorder sometimes considered psychiatry's heartland2 — schizophrenia. No other psychiatric condition has evoked such diverse opinions. Its biological roots have often been denied and, in the anti-psychiatry movement of the 1970s, there was even outright rejection of its existence. The latest paper, from the Schizophrenia Working Group of the Psychiatric Genomics Consortium, reports an analysis of more than 150,000 people and finds more than 100 genetic regions associated with schizophrenia, laying to rest forever the idea that genetics is not an important cause of the illness.

Dispute over schizophrenia's genetic basis has been ferocious. This is hardly surprising given that Swiss psychiatrist Ernst Rüdin — an early proponent of the argument that the condition is a single-gene disorder — advocated the view that people with mental illnesses should not have children3, and justified the sterilization and murder of people with schizophrenia. Despite the disease's long and contentious history, by the end of the twentieth century there was consensus that genetic factors were involved4. But agreeing that genetics has a part to play is not the same as finding individual genetic regions (loci) that contribute to disease susceptibility. Progress in this arena has been marked by false starts5,6, and by more than 800 genetic associations7 of dubious value.

Not unreasonably, many would ask why we should be any more confident that the consortium's geneticists have now got it right. The short answer is that the tests for associations between each gene in the human genome and disease are now mature. The correct criteria for determining significance in these tests are as familiar to human geneticists as their two-times table, and are exhaustively documented in the consortium's paper. Since the first genome-wide analyses for schizophrenia emerged in 2009 (refs 8, 9), the size of the studies has increased, as have the numbers of loci associated with the condition (Fig. 1).

Figure 1: Progress in identifying schizophrenia-associated genes.
figure 1

Since 2009, genome-wide analyses of the genetic loci associated with schizophrenia (including the current study by the Schizophrenia Working Group of the Psychiatric Genomics Consortium1, labelled 2014) have yielded increasing numbers of 'hits'1,9,15,16. Over this time, there has been a correlation between the number of people tested and the number of susceptibility-associated loci found.

The Psychiatric Genomics Consortium has now confirmed the existence of 108 loci that contribute to disease susceptibility. But even this number is insufficient to entirely explain the genetic causes of schizophrenia. So what does this discovery mean?

First and foremost, it confirms that genetics is a major cause of the illness. The risk variants now identified are common — they contribute in most, if not all, cases. This is a tremendous advance, of the sort that rewrites textbooks. Given the turbulent history of the field, this is a point that deserves emphasis and should be cause for well-earned celebration among those who carried out the work.

Second, we now know something about the probable genes involved. Readers not familiar with the vagaries of genome-wide association studies need to be aware that there is a difference between finding a genetic locus and finding a gene. Usually, it is assumed that the relevant gene is the one that lies nearest to the identified locus, but, as recently reported10 for a locus involved in obesity, this is not necessarily the case. As a rule of thumb, however, the closest gene is probably a good place to start.

So what genes are involved? Expectant readers will have to turn to Table 3 of the Supplementary Information (go.nature.com/koxm8i) to find all the details. Those who take the trouble will be rewarded with some delightful insights. For many years, a leading hypothesis has been that schizophrenia arises from abnormal neuronal signalling involving dopamine. This theory was based on evidence11 indicating that the efficacy of antipsychotic drugs correlates with their ability to block dopamine receptors. It turns out that those who championed the involvement of dopamine-related genes were right (a dopamine-receptor gene, DRD2, is associated). However, by no means all dopamine-related genes are involved (the gene encoding catechol-O-methyltransferase, an enzyme that metabolizes released dopamine, is not associated).

Having climbed a mountain of genetic analyses, it is perhaps a shame that the authors do not offer much of a view from the peak — no new biological hypotheses are presented. But it is clear that the idea that schizophrenia is merely a disorder of neuronal signalling (either dopamine- or glutamate-related) stands in need of revision. By far the most significant association is a locus on chromosome 6 that includes a region containing genes involved in acquired immunity: the major histocompatibility complex (MHC).

Given that the MHC also contains genes encoding various functional activities, the fact that the most significant association lies in the MHC is only suggestive of its role. However, Figure 2 in the paper contains another intriguing observation. The genetic effects in schizophrenia are enriched in regions outside the MHC that are also involved in acquired immunity. Is schizophrenia in part a disorder of acquired immunity? Surely this idea should start to be taken seriously. It is perhaps ironic that one of the earliest genetic associations reported12 was with the MHC.

It is possible that environmental exposure explains some of the genetic associations documented in this report. For instance, DNA-sequence variations in the CHRNA5-A3-B4 gene cluster are strongly associated with heavy smoking13. Smoking is extremely common (more than 80% prevalence) in people with schizophrenia14, and the consortium's identification of an associated variant in the CHRNA5-A3-B4 gene cluster might reflect a dose–response relationship between heaviness of smoking and schizophrenia risk, instead of — or as well as — a direct relationship. As ever-larger genome-wide association studies are conducted, we will presumably begin to pick up indirect effects. In other words, such studies may tell us about the environmental as well as the genetic causes of the disease.