Too Good To Be True? Debate about the Latest Schizophrenia Research
Treating mental illnesses like schizophrenia or autism, where symptoms range widely, would be much easier if doctors had diagnostic tests—tests that could tell them if two people who have overlapping but also different symptoms have the same disorder, or tests that could tell them which medications will work best for which person.
Recently, a study claiming that schizophrenia can be subdivided into at least eight genetic subtypes has garnered a lot of media attention. It’s easy to see why: if schizophrenia can be distilled into subtypes, each underpinned by different genetics, we might be closer to understanding the disorder and developing more personalized treatments.
However, the study, published online in the American Journal of Psychiatry on September 15, has been roundly criticized by others in the psychiatric genetics community, who fault the researchers for publically promoting results that others view as preliminary and in need of scrutiny and more study.
If true, the results support the theory that different people develop schizophrenia for different genetic reasons. There are lots of genetic suspects: most recently, a study published last July found over 100 places in the genome that increase risk a small bit for the disorder (read plain English and detailed discussion at the Schizophrenia Research Forum).
In this new study, Igor Zwir and C. Robert Cloninger at Washington University in St. Louis, Missouri, suggest that subsets of genetic variants occurring in the same person increase their risk for schizophrenia. Though harmless in and of themselves, some variants have been found more frequently in schizophrenia, and thus are identified as true, albeit weak, risk factors.
Using methods borrowed from computer science to detect patterns in large datasets, the researchers found groups of variants that increase risk for schizophrenia. The idea that genetic variants can add up to push someone over the threshold for a disease is not a new one, but the new study claims to be the first to identify precise groupings that matter for schizophrenia. On top of that, the distinct groups of variants occurred in people sharing similar symptom profiles. For example, one subtype included people who mainly experienced hallucinations and delusions, whereas another subtype was composed of those immobilized by profound apathy and disorganized thinking. Altogether, the authors report they have found eight genetically distinct types of schizophrenia (read plain English and detailed news stories at the Schizophrenia Research Forum).
Alas, the simplification may be too good to be true. Criticism of this study by other researchers was swift and strong. “We believe that the authors have not excluded important alternative explanations [for the specific groupings they identified] – if we are correct, then the major conclusions of this paper are invalidated,” they wrote in a comment published online in PubMed Commons. The critics identified five specific aspects of the study’s methods that they found troubling. (See links below.)
The controversy highlights the difficulties of making sense of the data streaming out of genetic studies of schizophrenia. The data sets are immense, and researchers are only beginning to explore how to make sense of them. Indeed, some insights may have to wait for yet-to-be-developed analysis methods. Zwir and Cloninger believe that their group has done that, but other researchers disagree. The authors or other researchers will try to repeat the experiment with another group of study.
This sort of debate is a necessary part of scientific progress and it happens all the time. We don’t see it very often, however, reflected in reporting by the general media. Headlines like “Schizophrenia Is Actually Eight Distinct Genetic Disorders” (Newsweek) are certain to grab the attention of many readers. What’s unfortunate is that usually only the preliminary, intriguing results get reported. We don’t generally read about the criticism of those results (as in this case, where the criticism has not been reported by mainstream media). And often no one is able to replicate results like these – that is, to get the same results in further studies. Science is messy, but it’s a trial-and-error process that has found cures and preventive measures for many diseases and disorders. We just have to be careful about believing everything that we read.
The paper itself (you can read a brief abstract without a subscription to the journal):
A few of the general media stories about the study:
The original criticism (posted by Gerome Breen) with responses by the study’s authors, Cloninger and Zwir:
Another version of the criticism:
Another story describing the controversy: