15 September 2014

Eight Schizophrenia Variants Subtyped

A new study has identified 42 clusters of genetic mutations that taken together predict one of eight sub-types of schizophrenia with distinct symptoms.  While individual genetic mutations (SNPs) are not very predictive of schizophrenia, genome wide associations with large sample sizes (4,200 people with schizophrenia and 3,800 controls) have demonstrated that clusters of multiple genetic mutations when found together are highly predictive schizophrenia genotypes.

The subtyping of symptoms, it turns out, was critical to finding the genetic clusters at all.  Even if a genotype is a match to one of the eight schizophrenia syndromes, it will show no significant correlation with the other seven and the signal that does exist may be diluted to the point where it cannot be seen.  In the same way, it is much easier to look for "tall statute genes" and "short stature genes" separately, than to look for "extreme stature" genes in a single statistical analysis.

Schizophrenia, in general, is about 80% heritable, but some genetic clusters predict that someone has a 95%-100% chance of having the associated particular subtype of schizophrenia, while other clusters predict that there is as little as a 70% chance of having schizophrenia and generally clarifies what type of schizophrenia that person who develop if they develop it at all.
In some patients with hallucinations or delusions, for example, the researchers matched distinct genetic features to patients' symptoms, demonstrating that specific genetic variations interacted to create a 95 percent certainty of schizophrenia. In another group, they found that disorganized speech and behavior were specifically associated with a set of DNA variations that carried a 100 percent risk of schizophrenia. . . it was only when the research team was able to organize the genetic variations and the patients' symptoms into groups that they could see that particular clusters of DNA variations acted together to cause specific types of symptoms.

Then they divided patients according to the type and severity of their symptoms, such as different types of hallucinations or delusions, and other symptoms, such as lack of initiative, problems organizing thoughts or a lack of connection between emotions and thoughts. The results indicated that those symptom profiles describe eight qualitatively distinct disorders based on underlying genetic conditions.

The investigators also replicated their findings in two additional DNA databases of people with schizophrenia, an indicator that identifying the gene variations that are working together is a valid avenue to explore for improving diagnosis and treatment. . . .

Eight classes of schizophrenia were identified by independently characterizing each phenotypic feature included in a genotypic-phenotypic relationship; classifying each item based on the symptoms as purely positive, purely negative, primarily positive, or primarily negative symptoms; and clustering these relationships based on their recoded phenotypic domain using non-negative matrix factorization. SNP sets harboring only positive symptoms are indicated in red, whereas those displaying negative symptoms are in green. Intermediate combinations including severe and/or moderate processes combined with positive and/or negative and/or disorganized symptoms were also color-coded.
Of course, while this allows for powerful diagnostic tools that could, for example, definitively confirm a claim of insanity by a criminal defendant who was not privileged enough to receive a clear diagnosis from professional mental health professionals before being arrested, this understanding does little directly to help someone who is diagnosed with a particular schizophrenia subtype cope with that condition.  But, someday, knowing a person's schizphrenia genotype might help someone know, for example, which type of anti-psychotic drugs, if any, are likely to be helpful for them.

No comments: