Schizophrenia not a single disease but multiple genetically distinct disorders | Hacker News Reader

Schizophrenia not a single disease but multiple genetically distinct disorders

news.wustl.edu

140 pointsMBCook70 comments

This is promising. If you read the DSM (the Diagnostic and Statistical Manual of Mental Disorders, published by the American Psychiatric Association), it reads a lot like a fuzzy pattern-matching algorithm:

    (A) Characteristic symptoms: Two (or more) of the following,
        each present for a significant portion of time during a
        1-month period (or less if successfully treated):

        (1) delusions
        (2) hallucinations
        (3) disorganized speech (eg. frequent derailment or incoherence)
        (4) grossly disorganized or catatonic behavior
        (5) negative symptoms, ie., affecting flatting, alogia,
            or avolition.

    Note: Only one Criterion A symptom is required if delusions
    are bizzare or hallucinations consist of a voice keeping a
    running commentary on the person's behavior or thoughts,
    or two or more voices conversing with each other.

What is a "delusion?" What is "bizarre?" There is a lot of subjective judgment here, and just generally speaking this feels more like alchemy than science.

I don't mean to criticize this approach too much though, because without a more detailed understanding of the underlying mechanisms and causes, this approach is the best possible way of at least clustering mental problems into groups. That at least allows people to share data about what treatments have worked for people in each cluster.

But if we can get a better understanding of the true underlying causes, and isolate the causes with better granularity, it should put us in a much better position to (I hope) provide much better and more specific treatment.

It's worth remembering that the DSM is not intended for use by the average person. It can and will use technical terms - like "delusion" - that users of the manual are expected to understand as technical terms.

Fair enough. It still seems safe to say that diagnoses based on criteria like "meets at least two of these five symptoms" are less than an exact science.

All models are wrong, but some are useful. This model has proven useful.

These symptoms do cluster. Most people never ever experience halluzinations, but then they also experience delusions or other psychotic problems, relatively specific social problems, paranoia, etc.

In some sense this is a human-readable equivalent of logistic regression. Because schizophrenia isn't as well understood as some other diseases, you can only make "exact" models about meaningless abstractions, or create fuzzy models which allow meaningful treatment decisions.

> All models are wrong, but some are useful. This model has proven useful.

The model is clinically useful, but the assumption (which the DSM does not itself advertise or support, which is made implicitly when its categories are used as the basis for research) that the diagnostic categories in the DSM represent commonality in origin is an impediment to research.

While many people have overstated the NIMH move away from DSM categories as a basis for research funding as something like an indictment of the DSM in its intended role or, even more hyperbolicly, a rejection of psychiatry as a field, it is an important move that recognize that the DSM categories' utility in their intended purpose does not imply that they represent real commonality.

Yes, exactly this. DSM criteria do predict effective treatment, negative consequences for the patients (both with and without treatment), and help plan for the costs of these conditions. These factors are clinically relevant.

Neuroscience tries to get a firmer grip on how the brain works, and what goes wrong when it doesn't. This research hardly touches the clinical practice as of yet, but ultimately diagnostics and treatment might improve largely.

All of this is exactly what I was meaning to say with my initial comment.

Current models are fuzzy, but in the absence of better information they are the best we can do, and they are useful.

Your points are well made. They're perhaps more applicable to Borderline Personality Disorder.

There are problems with having to match X of Y symptoms. Bob might be a close fit for 3 symptoms, and a weaker fit for 3 others, and not at all for the last 2. But, once Bob has the diagnosis, people interacting with Bob might not know what symptoms Bob matches and they might make assumptions - "People with your diagnosis do that because SYMPTOM".

It can be very frustrating.

Hmm ... what we mean by `exactness' - in physics, for instance - has been changing ever since the advent of quantum mechanics about a hundred years ago.

There are two points to consider: (a) the principle underlying the hypothesis is sound vs. is not, and (b) the computational model of the hypothesis is deterministic vs. probabilistic. These two have to be considered separately when performing an evaluation/verification.

What you have to recognize is that these critera are constructing a definition of the disease. It's not like the disease exists as some platonic object, and these criteria are simply the scientific way to test for it.

So where is the technical definition of "delusion" that the DSM is using?

In the glossary.

Really the issue is that psychology is in its infancy. Diseases that are well understood are defined and diagnosed by their root cause - the presence of a certain pathogen, a dysfunction of a certain organ. But when medicine doesn't understand the cause of a disease, psychological or otherwise, all that can be diagnosed is a cluster of correlated symptoms.

Take, for example, Irritable Bowel Syndrome - no one exactly knows what causes it, or if it's even a single disease, but there are a cluster of symptoms that indicate that a certain set of treatments might be useful (or not, depending on the patient). It is likely also a cluster of similar-looking diseases. Sounds a lot like the state of psychology, yeah?

Hopefully as we start understanding the brain, we can start diagnosing actual diseases with known causes, and not just a set of associated visible issues.

The DSM is hokey for sure. In fact, Thomas Insel, the director of the NIMH (one of the groups providing funding for this study) railed against it last year. The NIMH will not be funding studies based on DSM conditions. Rather:

> That is why NIMH will be re-orienting its research away from DSM categories. Going forward, we will be supporting research projects that look across current categories – or sub-divide current categories – to begin to develop a better system. What does this mean for applicants? Clinical trials might study all patients in a mood clinic rather than those meeting strict major depressive disorder criteria. Studies of biomarkers for “depression” might begin by looking across many disorders with anhedonia or emotional appraisal bias or psychomotor retardation to understand the circuitry underlying these symptoms. What does this mean for patients? We are committed to new and better treatments, but we feel this will only happen by developing a more precise diagnostic system. The best reason to develop RDoC is to seek better outcomes.

http://www.nimh.nih.gov/about/director/2013/transforming-dia...

I wonder if the results seen here are a result of this change in strategy?

There has been so much bullshit in psychiatry. My mother, who lived for her children, was essentially accused of child abuse by a psychiatrist during his explanation of my sister's diagnosis of schizophrenia to her. She apparently had been an over controlling and cold mother. None of her children would subscribe to that. However that accusation and the stress of caring for my sister over many years contributed, in my view, to my mother's untimely death to an autoimmune disease. I'm sure that psychiatrist probably never gave it any thought beyond that day. Psychiatry has a long way to go before I'll consider it has risen beyond quackery.

How old are you?

Those attitudes - that mental illness are caused by the family - are old and not acceptable any more. See RD Laing for an example of a psychiatrist who used to push this incorrect theory. http://en.wikipedia.org/wiki/R._D._Laing

> Those attitudes - that mental illness are caused by the family - are old and not acceptable any more.

"Not acceptable"? Your post suggests that, because the "refrigerator mom" idea has been debunked, that mental health professionals can't think that way any more. But this isn't how psychiatry and psychology work. An absence of solid science (of explanations) means that, even though a way of thinking has been discarded at the top, those at the bottom are free to keep thinking the same way.

This is why, more than a year after Asperger Syndrome was abandoned, psychiatrists and psychologists continue to apply the diagnosis. This is why, decades after homosexuality was dropped from the DSM, psychiatrists and psychologists are free to offer therapies meant to correct this "defect" in some of their clients.

> See RD Laing for an example of a psychiatrist who used to push this incorrect theory.

Incorrect or not, it wasn't a theory, defined in science as an idea about a cause, an explanation, with supporting evidence and falsifiable by new evidence. There aren't any of those in psychology.

But mental illnesses definitely CAN be caused by the family. A mental illness is the result of genetics and environment interacting, same as any illness. You can have a genetic predisposition to depression, but something in your experience of the world has to trigger it.

It's like breaking a stick. The stick can be really thin and easy to break, or really thick and sturdy. But either way, outside pressure has to be applied to break it. The amount of pressure required varies.

In general I agree with you - much of the mental health field is bullshit. However, there are some really good people who are trying to make it better. So, like much of medicine, the problem becomes trying to find one of the good ones...

It will take a generation or two for research like this to have any impact on practical diagnosis or treatment. The medical establishment is extremely conservative.

Consider the case of H. pylori and ulcers. This 2006 page from the CDC says, "Recently scientists have found that most ulcers are caused by an infection". (http://www.cdc.gov/ulcer/consumer.htm)

What is "recent" in medical research? Over two decades: http://www.cdc.gov/ulcer/history.htm

That is, the "recent" discovery of H. pyloi as the primary cause of ulcers was made in 1982, and general adoption of antibiotic treatment was still a work in progress in the late '90's and early 2000's.

On this basis, the current results--which will be difficult and expensive to replicate because the study is uniquely large and well-designed--will have at best a generational impact. Again, this is not entirely a bad thing: medical conservatism keeps doctors from doing a lot of harm, however frustrating it might be for those of us in the innovation business.

If you want to really cynical then it wasn't until the patent on Zantac expired that the H. pylori model made ground in the medical field.

Yeah, and the way these things are diagnosed is telling. (Basically, similar symptoms will tend to be classified as "borderline" or "schizophrenic" depending on the individual's demographics.)

yes, it's pretty ropey all in all - generally speaking people going to live with their families for a while has a higher success rate than anti-psychotic drugs, 60% vs 30%

in the 3rd world people can't afford drugs so tend to go for the family option

doctors in finland have adapted the 3rd world approach successfully and collected a lot of data along the way - incidence of schizophrenia is dropping in their region

http://www.mindfreedom.org/kb/mental-health-alternatives/fin...

I found this article a while ago that does a good job of detailing how psychiatry is absolutely not a science, and shouldn't be treated as such:

http://www.arachnoid.com/psychology/aspergers.php

This article does not seem to apply the same standards to itself that it purports to demand of psychology. Among other things, it freely makes false claims about the state of psychological research and then trots out a handful of examples to be assumed as exemplars.

It's little better than the Scientology "PSYCHIATRY IS DEATH" material.

Pot, meet kettle.

This article resonated with me very strongly, since I lived it. It seems to have struck a nerve for you too - maybe you're on the other side of the devout following?

Edit: I'm not claiming an emotional response to it, just that I can vouch for the validity of it in my case. Anecdotal evidence, but it's still a data point.

> This article resonated with me very strongly, since I lived it.

"X appeals to me emotionally because it conforms to and provides a pat explanation for my personal experience" does not imply "X is true".

In fact, disconnecting those two concepts is an essential step to beginning to approach the world scientifically.

You are incorrect. The article did not "strike a nerve" or otherwise provoke a strong emotional reaction in me. You may wish to stop and consider why you assumed that.

I'm glad to see advances being made. Schizophrenia is extremely hard to research. And it's devastating to the patient.

In the early nineties I worked at the University of Pennsylvania measuring brain MRIs as part of a schizophrenia study. They knew there was a strong genetic component, but they also knew that not everybody with a strong genetic weighting would develop the disorder. They had plenty of records of identical twins where one suffered from schizophrenia and the other did not.

The model then was the Diathesis-Stress model [1]. The diathesis, or predisposition, is in the genes. They thought the stress could be one of a number of things: maternal infection while the baby was still developing, birth complications such as brief interruptions in oxygen flow, infections after the baby is born, and perhaps extreme psychological trauma, like combat.

[1] http://en.wikipedia.org/wiki/Diathesis-stress_model

Most mental "disorders" seem to be just co-occurring negative symptoms aggregated by experts in a matter that is not arbitrary but also far from objective. The reason why e.g. anxiety and depression are typically comorbid might be because the practical distinctions between them are artificial and only exist to slap labels on nebulously defined mental ailments.

I believe that over time doctors will cease with the labels such as "inattentive ADHD", and merely treat debilitating inattentiveness based on the individual. The fact that ADHD can be hyperactive or inattentive, has always seemed bizarre to me. One person can't sit still, but my problem is I can sit still to a fault. But we have the same underlying disorder that's treated with the same medications.

The number of types of clinical depression, for example, is absurd. It's not merely depression, it's bipolar disorder. It's not merely bipolar disorder, it's bipolar disorder type II. Seriously? Decades from now we're going to regard this approach as highly primitive and bordering on pseudo-scientific.

Because every brain is different, trying to strictly classify these disorders would force people outside of a category, but the other side of that is that we have schizophrenia which seems incredibly broad "junk-drawer" sort of classification, to the point where it's easy to imagine someone who is highly eccentric being misdiagnosed.

I have no credentials obviously so this is all baseless conjecture.

I think that your criticism of the classifications is misguided. Certainly, depression and anxiety are different disorders, caused by different chemical mechanisms. And, bipolar is certainly different. And, I'm extremely glad that medicine has evolved and can now treat these diseases differently. With new dopamine agonists, sufferers are able to achieve normality. Unfortunately, there are a good portion of the population that suffer from this disease and others. And, 'categorizing' it helps both in diagnosis and treatment.

Biploar is different from depression, which is generally treated with SSRIs. I'm neither a chemist, or a molecular biologist, but I infer, from reading, that depression is caused by different chemistry, related to different parts of the brain. Depression and bipolar may have similar symptoms, at times, but the biological origins seem to be unrelated.

There's a strong body of evidence that bipolar is correlated with schizophrenia, autism, and ADHD. Again, this isn't my field, but each of these diseases relate to the dopamine pathways. And, there are quite a lot of genes that take part in and/or affect tyrosine-dopamine production. I also infer that the gradients of these diseases are caused by combinations of genes, in each individual. So, it makes sense that biploar I is treated differently than bipolar II. Once scientists have better understandings of genetics, treatments will become more granular and targeted. But, currently, if you're a sufferer, you must take a cocktail of drugs, in different doses, until you get the 'right one'. It's not the best way, but it's the best that we have today.

As an aside, I find the original article very surprising because it's widely known that multiple genes and SNPs are involved in incurring risks for developing schizophrenia (and bipolar and ADHD). To me, who continues to research the topic, I had just assumed that the results of the research were true. But, I'm glad that they were able to take controlled studies to help verify something that is deductive.

its scary when we talk about these disorders / behavior problems that we immediately go to prescribing a cocktail of drugs to 'cure', especially in light of ‘benzos’ being linked to alzheimers and such. drugs have their place, but i think they have over stepped it, where behavior modifications could do a lot more good with a lot less side effects like - destroying the brain.

> The number of types of clinical depression, for example, is absurd. It's not merely depression, it's bipolar disorder. It's not merely bipolar disorder, it's bipolar disorder type II. Seriously? Decades from now we're going to regard this approach as highly primitive and bordering on pseudo-scientific.

I doubt it. What we'll most likely do is view it as highly primitive in the same way that early developments in most fields are, and that the initial categorizations are systematizations of observations without even a strong hypothesis behind them. Its not pseudo-scientific, though, its part of what has to happent to start doing science. You have to define what the observed phenonmena are that you are trying to explain scientifically. (And, that being said, there is science that has been done based on it without, often, being heavily involved in exploring causes -- empirical exploration of how things in the category respond to environmental conditions is still science. We also have run into the fact that its often of limited utility and ambiguous results, likely because the categorization based on symptoms doesn't reflect the underlying mechanisms well, which is why we need more.)

Depression is not a single illness; it's likely to be several different illnesses that fall under a single umbrella diagnosis.

Please don't make the mistake of thinking bi-polar is in anyway similar to depression. They are very different illnesses with different treatments.

And, again, as we learn more about bi-polar we may find that it's a bunch of different illnesses that get given the umbrella diagnosis of bi-polar.

Your example of ADHD being treated with the same meds is just a function of our lack of knowledge of ADHD. So we might not always use the same meds for treatment.

In the future we'll have genetic testing to determine what meds will be useful or which meds are unlikely to work.

What? Depression is very similar to bipolar. Depression is a phase in most bipolar. The same meds often work for both, with mood stabilizers added for bipolar patients to try to arrest the manias.

That's not true.

You don't use the same meds for bi polar and depression - it is dangerous to do so.

People with depression should be given access to CBT with an anti depressant if needed.

People with BiPolar should not routinely take antidepressants, especially not if they are starting a manic phase. They should be taking lithium, valproate, or olanzipine.

http://www.nice.org.uk/guidance/CG38/chapter/Key-priorities-...

Huh. You're right. I guess the protocols have changed since I learned about them. Good to know!

You're treating the term "illness" as if they are actual objects, rather than just labels we give sets of symptoms and causes to make them convenient to discuss, research, and treat.

What?

I think you're agreeing with him?

This is potentially good news. My sister has schizophrenia, and not a day passes when I don't worry about my 3 year-old son being diagnosed with this terrible disease.

I'm not a scientist or an engineer, so maybe somebody who has read the published article can help me out here. I wonder what the percentages mean. How do they determine whether or not someone has a 73% chance of being diagnosed with the disease? And what does that mean exactly?

On the other hand, this research raises even more questions. For instance, I wonder what steps you could take to help a child with, say, a 60% chance of being diagnosed with schizophrenia? Are the genes like switches, in the sense that certain environmental variables can flip them on and off? If so, I guess there is a lot that could be done to help someone who has a good chance of being diagnosed.

This work investigates what are called "single nucleotide polymorphisms" (SNPs, pronounced "snips"), which identify genes that make slightly different forms of the same protein (the most common result of a SNP is to have swap one amino acid for another in a protein, which may change its shape a little so it won't work quite as well under some circumstances, or may bind a little more heavily to one site vs another.)

The percentages mean that if you take 100 people with a given combination of SNPs, 73 of them will get schizophrenia. People without that combination will have a much lower chance.

The good news for you and your son is that because they have identified multi-gene effects, the odds of you or your son having all the same SNPs as your sister is very low. You and your sister share half your DNA, so if there are ten genes involved, each with a 50/50 chance of being shared, that's a 1 in 1024 chance (2^10). For your son it's one in a million (unfortunately I can't tell from the article or the journal abstract how big the SNP sets involved actually are, but given the difficulty in finding them I'd expect more than four or five SNPs per set.)

With regard to your other questions, the honest answer is: nobody knows, although there will be any number of people trying to sell you on unproven solutions, mostly based on the demon of the moment. Twenty years ago it might have been red meat. Five years ago it would have been gluten. Today it would be refined sugars. Tomorrow it will be something else. The only thing we can be sure of is that the promoters will be smug, self-assured and self-deluded (or dishonest, but I think self-delusion is more common).

> unfortunately I can't tell from the article or the journal abstract how big the SNP sets involved actually are, but given the difficulty in finding them I'd expect more than four or five SNPs per set.

Table 1 ( https://pdf.yt/d/rErZm8HUluBbj8xu / https://dl.dropboxusercontent.com/u/5317066/2014-arnedo.pdf / https://www.sendspace.com/file/ht7shm ) list characteristics of the sets; they range from a low of 3 to a high of 32, with associated risks of 70 - 100% (!).

The size of the sets with a risk of schizophrenia between 90% and 100% is 11/11/19/4/6/3/10/6/9. (SNP set 87\_76 is the one with just 3 SNPs and a risk of 95%. Very nasty.)

Unfortunately I don't think this risk information is enough to to calculate how much predictive value SNPs would have for someone undiagnosed since you'd need to know how many healthy people have these SNP sets as well? But I'm sure the genomics hobbyists and anyone with 23andme data are probably looking closely at the paper and trying to extract risk estimates...

I would be very wary of putting number on the op child's risk of developing schizophrenia. You really need to know the frequency of the various snips in the population to know this. The most we can say right now is that the OP's son has an unknown higher risk than someone in the general population. The good news is we should soon have a better idea of what this risk is.

This is anecdotal, but having known schizophrenics, make sure your son stays away from drugs (e.g. marajuana and the like). I had a close friend who attributed his very bad schizophrenic symptoms to an increase in drug use (specifically marajuana). Correlation does not imply causation, but in this case, it's probably not worth the risk.

Narcotics will serve as catalysts for the symptoms. If they're present, they'll manifest themselves organically, regardless.

By "serve as catalysts" do you mean "make more severe," "make more frequent," "make appear earlier in life," or perhaps all of the above, or perhaps something different?

I think he means "make appear earlier in life".

> Several studies have suggested that regular, long-term cannabis use is one of a number of environmental factors that, in combination with certain genetic predisposing factors, may significantly increase a young individual’s chance of experiencing psychosis and developing schizophrenia. However, numbers of people being diagnosed with schizophrenia remained stable over time during which the number of cannabis users increased and average strength rose significantly.

http://www.drugscience.org.uk/blog/2013/10/24/think-cannabis...

What a coincidence. This was posted just as I was at the weekly "journal club" on human behavior genetics at the University of Minnesota with two researchers who used to be associated with Washington University in St. Louis (the source of the press release kindly submitted here). Genetic research on schizophrenia has been going on for a long time, and one of the WUSTL researchers I know locally[1] has been conducting that research since the year I was born. Experienced researchers on schizophrenia have always suspected that the finding announced in the press release headline is the correct description of reality, but it has taken a long time to do enough careful genetic research to be sure of multiple genotypes that all lead to the same clinical set of phenotypes that can be described as schizophrenia in current diagnostic categorization.

As other comments suggest, research along this line may eventually lead to more effective interventions for treating or even preventing schizophrenia. We already know (links below) that schizophrenia does have a very strong component of genetic risk, yet every once in a while identical (monozygotic) twins brought up in the same household are discordant for schizophrenia, so plainly some "environmental" risks matter too. It will be good to continue the research program on this devastating illness.

AFTER EDIT: Irving Gottesman is looking forward to reading the journal article by his former colleagues at WUSTL. Of course a preliminary finding like this will have to be replicated in other data sets to become part of established scientific knowledge. Then the hard work of matching treatments to patient genome patterns will begin.

[1] https://news.ycombinator.com/item?id=4661802

https://news.ycombinator.com/item?id=7663622

https://news.ycombinator.com/item?id=8069518

Fulltext: https://pdf.yt/d/rErZm8HUluBbj8xu / https://dl.dropboxusercontent.com/u/5317066/2014-arnedo.pdf / https://www.sendspace.com/file/ht7shm (excerpts: https://plus.google.com/103530621949492999968/posts/iJBeuxup... )

It's been interesting to see genetics slowly get expanded to help understand more conditions...

Just like cancer isn't caused by one thing, many of the common "psych" diseases aren't caused by one thing. The more obvious candidates for a strong genetic component are disorders like depression, alcoholism, and ADHD; but I'm sure this will be broadened into other mental health areas like diet, obesity and stress management.

There is still a strong environmental component, so I'm not trying to say that genetics is everything, but it's still important.

What's even more interesting is the insight this gives into biology. When we find a protein-encoding gene that, when mutated, causes a certain disease, it gives a pretty good clue as to that protein's function in the body.

Schizophrenia is more a catch-all term for "things are going severely wrong in a brain" than a disease.

I'll have to read the article to really understand what the authors have found out, but judging from the summary the advance is dividing the huge number of genes linked to schizophrenia into 8 groups of genes that associate or interact more closely than with other related genes.

In many ways, this isn't surprising. It's been evident that many genes have been linked to psychiatric disorders, but no "smoking guns" are evident, usually a gene accounts for <2% of the variance, and dozens or hundreds of genes have some influence.

So it's logical that it takes a cluster of many genes contributing something to have enough effect to be clinically apparent. How genes interact is an enormously complicated matter, and perhaps an even bigger question is the role of gene-environment interaction.

One thing to track is with the multitude of genes shown to be clustering, an important element is how "tight" the clustering is. I anticipate we'll find out as it's investigated further that overlap among clusters is significant, that is, "blends" or mixed/intermediate types are probably common. Certainly in the real world there are seldom pure cases of any condition, the boundaries tend to be fuzzy.

Not that the article implies at all that the new work gives "answers", but with any luck, it will lead to new insights, shed light on the astonishingly complex processes, and clues to making progress. Just expect the path will have many more twist and turns on the way there.

On a slightly different tack I found the 2010 Discover Magazine article about the role played by the HERV-W retrovirus in Schizophrenia interesting. Not quite sure how that line of research is panning out these days. http://discovermagazine.com/2010/jun/03-the-insanity-virus

My personal, uneducated guess about mental illness in general is that it's a systemic issue, having to do with a person's overall health (genetic + lifestyle/environment).

Systemic issues are likely hard to research, and so we take this scienctific (method) approach of studying small bits in isolation. This is similar to food science - there are too many factors to clearly study how we get nutrients from our food, and so science has studied bits in isolation (vitamin c, vs "orange juice with pulp and eggs").

My hope is that we'll find ways to study the overall health/status of an individual and see how that all adds up to behavior we deem "unhealthy", rather than being "trapped" poking at minute details (or perhaps well finally get enough detail to see the bigger picture).

Anyway I have zero expertise in this, but this is my guess on the situation.

That's really good news - if the same were discovered for bipolar and depression (which are often diagnosed along the lines of "well, you have these symptoms, so you probably have it") then it could help prevent a lot of misdiagnoses and increase quality of life.

That presupposes that a treatment exists for the "correct" diagnosis. In bipolar disorder there is a wide range of symptoms and prognosis. However, the subgroup don't inform treatment decisions to the point that the first choice of treatment usually works.

Any subdivision must lead to better treatment options and guidance to be meaningful.

I would say that there is no doubt that depression can be caused by a person's own beliefs about him/herself and the world. So it seems unlikely that such an explanation could account for all forms of depression.

The article doesn't mention it, but I'm curious if Watson involved in this (or similar efforts) at all.

No; the phenotype data was extracted from structured interviews:

"Ninety-three clinical features of schizophrenia from interviews based on the Diagnostic Interview for Genetic Studies (30), as well as the Best Estimate Diagnosis Code Sheet submitted GAIN/non-GAIN to dbGaP, were initially considered with the MGS sample (see references 31, 32; see also Appendix I in the online data supplement). The Diagnostic Interview for Genetic Studies was utilized for the Portuguese Island samples. Corresponding features were extracted in CATIE from the Positive and Negative Syndrome Scale, the Quality of Life Questionnaire, and the Structured Clinical Interview for DSM-IV (23). These phenotypic sets and their relations with one another characterize the phenotypic architecture of schizophrenia (Figure 1B)."

They did consider but reject using medical records (which is where a Watson might come into play):

"Despite the detailed phenotypic information we had available about subjects, there are still limitations to data obtained even from reliable structured interviews like the Diagnostic Interview for Genetic Studies. Interview data are based on self-reports that are interpreted and coded by interviewers. Subjects may not be willing or able to report their symptoms accurately. We had obtained information from treatment records and family history reports, but we chose not to use such additional information, except for the resulting best-estimate final DSM ratings of diagnosis, because its extent and quality varied in unmeasured ways between cases. The greatest limitation in the phenotypic assessments in available GWAS databases has been the overreliance on subjective symptoms with an absence of objective measurements, such as cognitive tests, brain electrophysiology, and neuroimaging (37). Subjective symptoms are fuzzy indicators of the underlying pathophysiology. Objective measures could complement the assessment of symptoms and could be applied to both cases and controls, thereby providing a more comprehensive and valid characterization of the phenotype of all subjects. The biggest challenge in GWAS is access to studies with rich phenotypic data about both subjective and objective measures obtained systematically from all subjects."

I sure hope not. Associating genetic variations with symptoms or physiological findings is a relatively clear-cut statistical problem. The major challenge in whole genome association studies (or their extension to whole genome sequencing data) is to get enough data to draw meaningful conclusions.

There is a cognitive dissonance in biology/medicine: On the one hand they complain about too much data to process, but at publication time, the power is too low...

The problem with a lot of the data in electronic medical records is that they are not structured but in natural language. This is exactly where Watson can help.

I don't think so. It is too expensive to enroll more than a couple of thousand of participants for a GWA study, though they are often pooled for a broader analysis. Anyway, genotyping will probably cost at least $100 (depending on the technology a lot more), which puts manual entry of patient data into perspective.

An unpaid intern can be cheaper than fancy artificial intelligence!

The idea that schizophrenia is a collection of different diseases is not new at all. When I did my neuropsych degree in the early 90s, the lecturers commented that schizophrenia was likely to be at least four different things (this article suggests eight). Their commentary was that schizophrenia was a sort of dumping ground where people got put if they didn't fit into a different diagnosis, and I remember the run-down of frequency of symptoms - the most common symptom exhibited by schizophrenics, auditory hallucinations, was only seen in 69% of patients.

The idea of the four broad groups at the time (from very rusty memory here) stemmed from plus/minus showing cognitive/perceptual disturbances (like auditory hallucinations) and plus/minus showing social withdrawal symptoms, along with a couple of other things I can't recall.

It's not to cast aspersions on the article - mental health research is always good - just to say that this idea has been bouncing around for quite a while.

I find this research to be very interesting. It should make it much easier to pin down the cause of the individual features of Schizophrenia, many of which are believed to be caused or triggered by the modern Human diet (refined sugars, grains, other (sometimes highly) processed foods et al.) by a growing number in the scientific community.

> many of which are believed to be caused or triggered by the modern Human diet (refined sugars, grains, other (sometimes highly) processed foods et al.) by a growing number in the scientific community.

Links to studies published in peer-reviewed journals or GTFO. "Modern diet" is the medieval humorism of the last decade.

Lots of links here.

http://evolutionarypsychiatry.blogspot.com/2013/11/gluten-an...

There is also rising evidence that gut flora correlate with various mental health issues.

http://www.apa.org/monitor/2012/09/gut-feeling.aspx

http://www.scientificamerican.com/article/gut-bacteria-may-e...

From your first link, the only study that attempted to show a causal link between gluten and schizophrenia had no control group, no placebo, and n=2. This isn't a study, it's a guess.

There may well be something to this whole thing, but the science just isn't there right now, at all. Right now what we need are studies, not solutions. The evidence is non-existent.

Sounds like malarky to me. Sounds like a group to understand something outside their reach; finding subjective data points to further categorize the problems allowing themselves to believe they now have more of a handle on than they did before. I don't believe they know anything that adds more value.

Reply on news.ycombinator.com