Hidden Valley Road - Inside the Mind of an American Family - Robert Kolker Page 0,127

was finding potentially relevant genetic locations all over the place, suggesting a new and deeper understanding of how mental illness operated in the brain. In the years to come, this new knowledge would help geneticists see how schizophrenia and other mental illnesses significantly correlated with copying errors—or copy number variations (CNVs)—in which whole chunks of DNA are either overproduced or go missing altogether. But for those hoping that the GWAS approach would find just a few genes to pin the blame on, this was hardly encouraging. And this was just the beginning. The schizophrenia GWASes that followed identified the first several genetic locations that seemed especially relevant to the illness. One GWAS, published in Nature Genetics in 2013, included about 21,000 genetic samples and found 22 such locations. Another GWAS, published in Nature in 2014, involved 36,989 patients and found 108 locations. Robert Freedman’s CHRNA7 gene was in one of these suspect locations, which offered him some nice outside validation. But the more they found, the less meaningful the results seemed.

Each of these genetic irregularities, taken by itself, only accounted for a minuscule increased chance of an individual having schizophrenia. The researchers tried to make lemonade out of lemons by considering all of these insignificant factors together, combining them to come up with what they touted as a “polygenic risk score.” But to many researchers, the polygenic risk score was merely a lumping together of trivialities into something only slightly less trivial. The genetic markers identified in the 2014 Nature GWAS, taken together, would only increase one’s chances of having the disease by about 4 percent. “It’s sort of a mindless score,” said Elliot Gershon, DeLisi’s old boss at NIMH, who had moved on to the University of Chicago a few years after DeLisi left. “You can’t really tell anything from the polygenic risk factor.”

The GWAS approach was not delivering the tidy ending that geneticists like Varmus had expected. In the face of blistering disappointment, the leaders of the Broad Institute, which had been leading the GWAS efforts for schizophrenia, decided to double down—resolving to build a bigger and better GWAS. “The guess among my colleagues is that we’ll need 250,000 schizophrenia patients,” said Steven Hyman, the head of the Broad Institute’s Stanley Center for Psychiatric Research, “which is daunting, but feasible for this disease.” By the time they were done, Hyman predicted, “there will be thousands of variants in many hundreds of genes” all pointing toward schizophrenia.

Some suspected that the entire process might be leading the field astray—sending researchers once again to go look for their lost keys where the light was, not where the keys really might be. After all that work, the underlying nature of schizophrenia remained a matter of intense debate. “Is it a classical organically based biomedical disorder,” the psychiatric geneticist Kenneth Kendler wondered in 2015—like, say, Alzheimer’s disease was thought to be—“or is it the severe end of a spectrum of syndromes that aggregate together in families?”

Lynn DeLisi knew where she stood on the matter. She’d known for years. “My thought was, ‘I don’t believe that these hundred genes or markers are going to lead to anything,’?” she said. “I want to see what’s causing schizophrenia in these large families like the Galvin family.”

* * *

—

WHEN PFIZER PULLED the plug on DeLisi’s research into families with schizophrenia in 2000, she was forced to stop all of her work. Like the parties in a divorce, she and Pfizer divided her physical samples from her families straight down the middle. The term in research is “aliquoted”: She and Pfizer each walked away with one half of each of her blood samples, enough material, in theory, for both parties to continue work. But in a cruel irony, no work would continue: DeLisi had the will to keep going but not the money, while Pfizer had the money but lacked the will.

Why would any large pharmaceutical company not want to try to develop a better drug for schizophrenia—one that might hit a genetic target and resolve issues that Thorazine and its offshoots never could touch? The reasoning at the time, according to professionals dealing with those companies, was pretty clear. Even with a genetic target, like Freedman’s α7 receptor, the pipeline to develop and test such a drug was extremely expensive, requiring human subjects willing to endure unpredictable side effects. Which would be