Creating a Cellular Map of the Human Brain

“We don’t know how many different cell types there are in the brain or how best to define them.” Michael Oldham, PhD
 

Fifteen years ago, Michael Oldham, PhD, was leading a team of 10 at a digital advertising agency. Feeling adrift and unfulfilled, he took a step back to reflect on the things he’d most enjoyed in life and in school. That period of self-inquiry led to a realization that he was passionate about science, and more specifically, driven by a desire to understand fundamental differences between human and chimpanzee brains. Today, Dr. Oldham is an assistant professor in UCSF’s Department of Neurological Surgery, leading an effort to create a cellular map of the normal human brain. That map has the potential to serve as the foundation for a host of future diagnostic tests, allowing for the comparison of a diseased profile against a normal profile to pinpoint biomarkers of neurologic and psychiatric diseases.   

How he began this work: “As a graduate student, I looked for patterns in the human brain transcriptome, the set of all genes expressed within a given tissue sample. I started seeing the same patterns over and over again, and I realized that these patterns were telling us a great deal about the underlying cellular organization and molecular basis of cellular identity in the human brain.”

A new approach to analyzing gene expression data: “It was kind of a crazy time because I was a graduate student, and I realized that almost everyone under the sun was analyzing gene expression data by looking at one gene at a time. As it turns out, that kind of approach completely ignores all of these co-expression relationships that exist in the data. Effectively, there is a lot of information that’s invisible to the standard method. My big insight was that many of the recurrent gene expression patterns I saw in the brain were driven by variation in the abundance of different cell types. But even today, we don’t really know how many different cell types there are or how best to define them.”

A productive first year: “I spent my first year at UCSF holed up in my office, writing code and building a computational pipeline for analyzing gene expression data. That turned out to be a great investment, because that code allows my lab to crank through these giant datasets in a rapid, systematic, and standardized way so we can catalog the patterns we see, where we see them, and what we think they mean. A lot of what we do is reanalyzing other people’s published data because so much has been generated, but people have just scratched the surface of what’s there. I can pull the data down, aggregate it, run it through the machine, and extract patterns.”

How he’s expanding this work with the UCSF Weill Award: In building this map of normal human brains, I’m hoping to quantitatively define the molecular basis of cellular identity in the healthy human brain. With support from the Weill Award, I’ve been able to add another dimension to this project, and I’m now working to also quantify the abundance of cell types in the same tissue samples.”

Why this work is so crucial: Believe it or not, if you take a given volume of brain tissue from a normal person – someone who’s died but has no evidence of neuropathology – there are no established ranges of normal variation for the abundance of different cell types. In other words, it could very well be that schizophrenia is caused by a deficit of a particular type of interneuron in a particular region of the brain, but we don’t know what normal looks like, so how can we find it?”

An outsider’s view: “A lot of people have underestimated what you can learn from studying human brain samples and are focused on mice instead. But there’s a huge amount that you can learn from studying even a single human brain specimen. I believe that if you’re studying human diseases, you need to study human tissue first and foremost. In neuroscience, that scientific worldview has been surprisingly uncommon, but times are changing.”