A cancer physician and researcher, Dr. Mukherjee won the Pulitzer Prize for his book, The Emperor of All Maladies: A Biography of Cancer. We’re reading his latest, The Gene: An Intimate History (and highly recommend it).

1. We are interested in the future of technology and human experience—we feel that your book The Gene has a lot to say about this. As we get closer to mapping our own genetic futures, do you think we’re all pre-survivors of a trans-human utopia, or a dystopia?

As we’ve started to unleash powerful capacities to interrogate genes, particularly using deep-learning tools, I think there is a sense that there will be many more pre-survivors based on genetic predictions. And by this I mean that more of us will become survivors of diseases we haven’t yet had.

2. Can you talk a bit more about the intersection of deep learning and gene-editing?

How CRISPR/Cas9 gene-editing technology intersects with deep-learning is really quite an open question. Right now, CRISPR/Cas9 is a precision tool for making, usually, changes in single genes, one gene at a time. That’s what its advantage is—it’s an exquisite precision technology that allows you to make direct changes to the genome. The arena where machine learning is immediately intersecting with all of this is not in writing or changing the genome—but in reading the genome, and in explaining to us how genetic variation can contribute to human variation, including variation in disease.

3. In terms of the human genome, what does it mean to have intellectual property? Is the data we get from 23andMe genetic IP, so to speak?

The courts have had an evolving understanding of this. Generally speaking, the courts in the United States have thought about genetic sequence information as parts of the body, or parts of nature. And in general, IP around parts of the body or parts of nature has been hard to do from the IP perspective. But an algorithm which creates the pattern out of genomic information is not a part of nature—it is a manipulation of nature. So those algorithms can very well be proprietary. Now, the separate question is, what’s the value of that? What’s the commercial value of this?

4. When CRISPR hits it prime, do you foresee there being some kind of recipe for desirability that one could follow as a germline intervention early on?

For powerful single-gene mutations, I can conceive of this in the future. You can imagine that you’d not want to bear a child carrying a mutation that would eventually cause some unfortunate disease with extraordinary future suffering. That part lies in the realm of imagination. But what doesn’t lie in the realm of the imagination is what to do with most human diseases, which will end up having not 1-2 gene variances that contribute to it, but hundreds. CRISPR is very good at making deliberate intentional changes in single genes, and therefore its very good in setting up a kind of future scenario where single genes can be intervened on in the germline or in the body. What we’re finding out more and more from human biology, and this is important, is that although there are certainly cases where single-gene mutations are responsible for a disease, the actual people or most people have diseases that are manifestations of hundreds, if not thousands, of gene variants. That, and the environment.

5. There are potentially unseen advantages to seemingly negative genetic predispositions: you’ve used the example that creative people may be predisposed to being bipolar. If desirability is a moving target, and evolution changes according to culture and environment, what is the dream, in your opinion, for gene editing?

This is hard to answer because I really think of gene editing in three buckets—and it’s very important to distinguish between the three, as they have different ethical and technical challenges. The first one, which is happening now, is genetically modified organisms—making different kinds of crops, and changing gene drives in pests, in fish, in animals. The dangers here are obviously ecological—the main concern is having humility about ecology, and how little we know about interrelatedness in ecology. Category 2 is using CRISPR/Cas9 to make genetic changes in human cells, but not in sperm, eggs, or embryo cells. This is called Somatic Gene Editing. We are doing this, my own laboratory does it. We have been shocked by how easy it is, and the efficiency, though we still have questions about whether there are off-target effects. And there are social and political challenges like what disease to apply it to; there are trials in the next 2-3 years that will tell us this. The third and final category is the most ethically complex, and the most technically complex—and it’s also the one that has everyone fired up. And that is making these changes in sperm cells and eggs cells, or cells that make sperm and eggs—so-called germline gene editing. I think we will begin to see some early data from germline editing in the next 5-10 years. It’s the tallest order in terms of what the challenges are and how much we would be able to see in the next ten-odd years.

This interview was originally published in Tech Fancy Issue 3: Un/Natural Selection.