Scientist of the year, R & D 100 Magazine

Mosaics: Recollections of a scientific endeavor.
Note: There is some scientific jargon in this remembering; if you not a scientist in my field, but curious about the vaccine, please excuse me and just skip those bits.

This award was a lovely surprise. It took a "village", lots of folks many years, get this vaccine into clinical trials. So here is some history if anybody is interested, of those who moved this concept along. I was initially inspired to try this idea based on the global database that my team at LANL keeps -- HIV diversity is an incredibly challenging problem for a vaccine, and minding the HIV database convinced me we needed to try to address the diversity directly, by trying to "see" the virus the way the immune system does. I wanted to design a vaccine that could trigger immune responses by allowing recognition of the most common forms of an epitope simultaneously, and cross react with diverse HIV strains (an epitope is a small bit of viral protein that an immune response recognizes). I also wanted to leave HIV proteins intact in the vaccine, so they would be easy to deliver and be processed naturally, and so antibodies that recognized folded proteins might be elicited by the vaccine as well as T cells. I came up an approach to the problem, evolving sets of protein sequences on a computer, that in combination could potentially optimize immunological coverage. The idea was partly inspired by the way HIV evolves in nature, partly by a machine learning concept called a genetic algorithm. It took me a couple of tries at a grant to get it funded; the first reviewers thought the idea was completely nuts, that this kind of thinking could never EVER work. I eventually got funded to bring together the scientists I needed to do the "theory" part of the project through an internal LANL internal grant which fostered exploratory research (a DR). A wonderfully talented machine learning computational scientist, Simon Perkins, was on the team, and with input from my also wonderfully talented husband James Theiler, Simon fine tuned the algorithm, got it in a good computational framework, and coded it up.

I tried to get my old friend Satish Pillai to come work on this project with me as a postdoc, but he was busy with a new job at UCSF, and even green chili couldn't get him back to New Mexico. (See, Satish, I _told_ you it was going to be a good project!) Satish of course now has his own lab, and 2018 found us still working together -- a bit on music, and a bit on an HIV therapeutic strategy, tugging on some crispr threads. But back in the day, Satish was a no, and so I hired Will Fischer instead, an unknown at the time, but a man who very clearly loved to think about evolution, so the project was a very good fit. Now Will is a long time friend and colleague, but then he was a newly minted postdoc. He helped me put Simon's code through its paces, and we explored together how it performed "in theory", working out ways to visualize the comparisons we needed to make (James also helped a lot with this part), and how to best inform the design of the actual vaccines for testing: global versus regional, one, two, three or four proteins sets, which proteins to include. Karina Yusim, Carla Kuiken, and Tanmoy Bhattacharya also helped us think about this. Mosaic code does not require a fancy computer, and we never used supercomputers on the mosaic project. Still this was all tricky business, with multiple false starts and restarts as we were getting going (like initially trying to add in HLA, then realizing, partly inspired by Steve Self, and partly simply by the very high density of epitopes in the database, that we could just treat every epitope-sized bit of the protein as a potential epitope and leave aside HLA complexity and frequencies). Once we had the basic code, we needed to address parameter optimization, tweaks were added to make the code run faster, and we made gradual improvements in output and interpretation. Our goal was resolving what might matter biologically as well as we could based on information in the database. For me it was several years of very very long days, and this comes with a high life-cost. And it wasn't trivial; things always seem more obvious in retrospect than when you are in trenches making something new.

As we worked, I regularly talked through the progress of the theory, for over a year, with my dear friends who were experimentalist colleagues, Bart Haynes, Beatrice Hahn, and Normal Letvin (now passed away), getting their input and advice. They shared good ideas. The four of us had gotten some NIH funding so we could turn the mosaic design theory into some experiments to test it. Bart and Norm did some initial studies, and Bart continues to work on mosaic related vaccine concepts with me currently, and we are working on second and third generation developments. We shared the mosaic concept with Dan Barouch in early days, and we made him some global designs, and happily Dan ran with them. Dan wanted us to only have two mosaic components per protein; given that constraint, we tried to design something with a chance at being able to work all over the world, Africa, US, Europe, Asia... It was Dan's experiments over many years that lead to the current clinical trial, and it is Dan's "vector" that delivers the HIV mosaic vaccine. The mosaic would not be in human efficacy trials now without his many ideas about good ways to express the vaccine insert, and his thoughtful and careful preclinical experimental work, and his collaborative spirit bringing in an industry to help take it forward.

We won't know if this vaccine will protect people from HIV for another couple of years; so this "scientist of the year" is a bit premature, though I'm really happy and honored that it happened. On the hopeful side the mosaic vaccine does significantly slow down infection of virus exposed monkeys, and if vaccinated monkeys finally do become infected after multiple exposures, they stay healthier and don't progress to AIDs. We don't know if this will translate to preventing infection in people who will be confronted with HIV in all its natural diversity "in the wild", or if it will help them fare better if they are infected. We cannot know that until it is tested, and it takes time. But I harbor hope. If it doesn't protect people, still we will learn some things from trying.

Meanwhile, there are many other good HIV vaccine ideas in the pipeline -- and a few of them are my own! Its why I can't retire yet, I want to see the ones with promise through.

Thanks to everybody who has sent kind notes and good wishes.

I'm keeping some pretty classy company in this award, the nerds among you will know some of the names of past recipients:

Dr. William Pickering 1967 JPL
Dr. Wernher von Braun 1969 NASA MSFC
William Lear 1971 LearJet
Dr. Mary Good 1982 DOC
Dr. Justin Rattner 1989 Intel
Dr. Kary Mullis 1991 Nobel Laureate
Dr. Susan Solomon 1992 NOAA Ozone Hole
Dr. Tim-Berners Lee 1996 WWW
Dr. J. Craig Venter 1998 Human Genome Project
Dr. Eric Lander 2003 MIT/Whitehead Institute
Dr. Anthony Faucci 2005 NIH
Dr. Steven Chu 2011 DOE
Dr. Robert Langer 2012 MIT
Dr. James Tour 2013 Rice Univ.
Dr. Karl Deisseroth 2014 Stanford Univ.
Scott Kelly 2015 NASA
Dr. Dharmendra Modha 2016 IBM
Dr. Cori Bargmann 2017 Chan Zuckerberg Initiative