What do Sir Isaac Newton, Oppenheimer and a Hedge Fund Billionaire Have in Common? Molecular Dynamic Simulations in the Chong Lab, of Course.
Every now and then, a tree will become infested with mites or infected with fungi, and out of the trunk will grow a cluster of twigs that look like the bristles of a broomstick. It’s called Witches Broom Disease, ostensibly named such by more superstitious forebears.
A healthy red oak tree will have, in terms of size, a natural derivation of branches. Perhaps the trunk splits into two branches, and those branches keep branching out to smaller ones and finally to twigs holding oak leaves. An oak tree with witches broom disease could have a trunk bristled with tiny shoots.
I don’t think Isaac Newton stood on the shoulders of giants. He climbed onto a mountain — or a tree — of little people. Standing on giants’ shoulders would be akin to a twig growing out of a tree trunk. Newton was a leaf on a twig on a stick on a small branch on a large branch on a giant trunk. That is healthy research and he most likely would not have been the great scientist that he was if he had been a witches’ broom sticking out of the giant trunk of Science.
I have known research is collaborative. I know I am not saying anything new by saying research is evolutionary. It might be new to say research is relational, intimate even. Then again, it might not be. But it’s a privilege to have seen it. I had the privilege this summer when I became part of Lillian Chong’s computational biophysics lab-family. Interviewing one of the former lab members, Matthew Zwier now a professor in physical chemistry at Drake University, he said, “The truth is always simple. It’s the process of finding it that really isn’t. Research is messy. There’s a trial and error process and backtracking and some seriously complicated toil to find the simple truth.”
-
The Chong lab does not look like your typical chemistry lab. There are no boiling-flasks or colorful solutions bubbling through reflux columns. It looks more like an internet café. In exchange for fume hoods, several desktops line the room. A couch sits next to the entrance, where Ali, a grad student in the lab, maintains his post. On the far wall is a high bookshelf with a range of subjects from organic chemistry to C++ programming. In front of the bookshelf is a large circular table with some snacks on it. A coffee maker on the windowsill percolates and the smell of an Arabica blend stimulates the air. Adam and I sat with Corinn at her computer by the windows of the room.
“Scroll over — see that line going over there?” I said, pointing off the computer screen.
Corinn scrolled over and up.
“Vijay Pande. That was Lillian’s Post-Doc advisor at Stanford,” Adam said.
“Bruce Tidor on the other end — her undergraduate advisor at MIT.”
“Oh, there’s Martin Karplus,” Corinn said pointing at his little name box two above Tidor and amidst the large web of names.
“Yeah,” Adam said. “He, along with two others sort of invented MD. Got the Nobel Prize in Chemistry in ’13.”
“What’s MD?” I asked.
“Molecular Dynamics.”
“Oh, of course.” I felt a little red in the face. Molecular dynamics (or MD) was everything in the lab. Now early June, I had been there for about three weeks and had thought I would have heard the acronym a few more times. MD is a rendering and simulation of molecular events and processes. It is not uncommon to hear of an experiment done in silico which is, as one might guess, an experiment done on a computer. The result is essentially a movie. Biological processes too small and too fast for the naked eye, or even a microscope, are simulated and the result is a video in which it is possible to see every last atom as it moves in space. Adam, Corinn, and I were sitting in the lab in Eberly Hall, browsing through the MD predecessors and successors — Lillian’s Ph.D. ancestry. The tree was large and, gawking at all the names, I had a feeling of both reverence and disrespect. It was like reading the names on an obituary; you read a name and may wonder or even know a little about them but you know that knowing their name or who they are related to really does them no justice as to who they actually were.
“Oh, ok,” — I pointed over at the upper left — “Ok, right there. Niels Bohr. Who was he again? His name sounds familiar.”
“Scientist,” Corinn said.
“Yeah I had heard Lillian was related to him somehow,” Adam said.
“And there’s Lennard-Jones. Like Lennard-Jones Potential,” Corinn said.
I saw Peter Kollman — Lillian’s advisor at UCSF. I didn’t know much about him but I remembered that Lillian had spoken quite fondly of him.
At the upper right hand of the monitor was a little scroll down tab that said: Tree Length.
“Wait, click on that,” I said.
Corinn hit the tab. A dropdown appeared with a list: four lines, five lines, ten lines.
“Click ten.”
The tree stayed the same size in the screen but the scroll bars grew smaller. The tree had grown and it took many screen frames to fit the tree inside, with lines sprawling out to sides and going up and up to older and older chemists.
“There’s Linus Pauling.”
“That’s amazing,” I said, “One day, people will be scrolling through this ancestry and will be shocked to find they are related to you two.”
Corinn threw her head back and laughed. “Oh right, we will see about that.”
“Don’t you think Karplus and Bohr might have laughed too?”
“Who knows? Let’s check out this guy,” said Corinn clicking on a random name on the side search. The screen loaded and then his name popped up in the tree field. We waited for his tree to load — and then realized it had. We looked at each other and then broke out with laughter.
“Oh, he must have granted himself his own Ph.D.” said Corinn.
“Oh yes,” Adam said, “No advisor was good enough to be his. He had to become his own.”
“Do you think he gave himself a hard time on his dissertation?” I asked keeping a straight face. “I wonder how many times he had to defend.”
We returned to Lillian’s Ph.D. tree and perused through some more scientists: people I had never met nor knew. But ancestors of people I did know. Ancestors of Corinn and Adam. Ancestors of Ali who sat pensively on the couch as usual, teaching his young padawan, Gabe, in the ways of brute force. Ancestors of Matt who worked with Adam to develop the first binding pathway simulation of the tumor suppressor protein, p53 with the protein MDM2; a crucial biological pathway in the first stages of cancer.
And of Lillian, who managed her lab not with a heavy hand but with grace. Corinn and I would talk about her occasionally, quipping to each other that no one was as good as her. It was then, when we joined her lab, that we knew it was better to be loved than feared. Lab members could come and go as they pleased during the day as long as they were making satisfactory progress. And I don’t believe I had seen researchers with more ambition and a harder work ethic than those in Lillian’s lab. Lillian advised and criticized work but never in a demeaning or hurtful way. It was always encouragement and she simply emanated a positive atmosphere wherever she went.
These were the ancestors of Karl, too. Now living in Wisconsin, I interviewed him on Google Hangouts about the work he did in the lab for force field development.
“Where did you grow up?” I asked rather abruptly. It was my first interview and I hadn’t learned the art of interviewing through conversation rather than rapid-fire questioning.
“Reading, Pennsylvania” Karl said.
“What were you doing as a kid? Like hobby-wise?”
“Playing video games.”
I had heard, and it did seem, that he was pretty private, or at least disinterested in answering personal questions. Lillian told me that one Monday morning a few years ago, he walked into the office and announced to the lab that he had gotten married. It turned out that many did not even know he was dating.
But once I got into the science with him, it was hard to keep up.
“Could you tell me what exactly this ‘failure’ was that started your research?”
“It was a correction for the salt bridge interactions. They had to be rewritten because we found out that the parameters for the electrostatic force field were way stronger than their experimental values. We used a variety of amino acid side chains to try the fit the new parameters to experimental values.”
“So you’re saying the salt bridges were bonded too strongly?”
“Yeah. The charges’ coulombic interaction was modelled as two electrostatic point charges. It turned out that was an oversimplification and one that made it bond too strongly, so we had to model for polarization of electron density rather than simple electrostatics.”
“You said that all the force fields were Newtonian. With a polarization force field, this is obviously no longer true.”
“Well it still uses Newton’s first law, F = ma. But yes, polarization is a quantum phenomenon. Newton’s forces, in this case Coulombic forces, were no longer accurate enough.”
I spoke with Karl for about an hour and a half, learning all that I could wrap my head around and writing down verbatim whatever I couldn’t. Over the course of Karl’s work, the AMBER protein models had undergone two new developments from AMBER ff13ipq to 15. The AMBER models are software through which proteins are rendered in computer space. They are the code which yield the proteins for simulation. Developments of these models are merely improvements of this rendering software; they make the models more like proteins as they occur in the real world, in our bodies.
“There’s sort of this phylogeny of models — kind of like life. AMBER ff13ipq went on to yield ff14ipq. And then some work went in to yield ff15ipq from ff14ipq. That’s a lineage and it keeps branching out until two random protein models with a common protein model parent have little in common.”
“So that’s an evolution, if you will, of the bonded parameters. How about the non-bonded?”
“Many of those stay the same. Obviously, Coulomb’s law was attenuated. Most of the same code for them go back to the 90s, I think. They’re much harder to develop — it kind of ties all those AMBER models together.”
At times, Karl would let loose, clearly pleased to discuss and dissect the subject of force field re-parameterization. He would get into the weeds of the parameters and code. A lot of it would go over my head. I realized during the interview that he had not yet mentioned a single other person about his research. And I started seeing him as the eccentric, lone, mad genius whose un-relatable acumen had and needed no peer.
I was too curious. “Did you have any help while you were developing these force fields?”
He looked at me, blankly. Then he laughed. “Oh, I had plenty. Many people were involved in this project but you could say that the bulk of the work was pretty evenly split between me and Dave.”
I had heard about Dave in the lab a few times at this point. Dave lived out in New Jersey, currently doing research at Rutgers. He never worked in Lillian’s lab but he collaborated with Karl, developing the AMBER protein models. From my understanding, Dave was the other side of the coin, as it were.
So about a month later, in early July, I logged on to Google hangouts and talked to Dave. “So I was wondering if you could tell me a little about where you grew up,” I said.
“I grew up in Los Alamos amongst a bunch of smart, rich people,” Dave said wryly. “I mean, you could probably guess why. It’s Los Alamos. I’m sure you’re familiar with the Manhattan Project. My dad ran a tight ship working on bomb simulations in a theoretical computational lab. These simulations were essential before the real detonation so that the bomb’s explosive power could be predicted. Needless to say, it’s important to have an idea, and an accurate one at that, of how much force and how many tons of TNT of equivalent energy these bombs will release before detonation. And it’s interesting too…a lot of the modeling for the bomb explosions involved the use of fluid dynamics. After all, metal becomes liquid when subject to those incredibly high temperatures.”
At this point, I couldn’t help but remember Karl’s interview, particularly when I asked him about where he grew up, to which he simply replied, “Reading, Pennsylvania.” It would be unfair to call Karl close-mouthed and Dave, motor-mouthed. But I did begin to see differences in them, complementary in fact. Dave appeared to be the visionary; he was the guy who seemed to know a bit of everything about everything (though by no means a jack-of-all-trades; he is a specialist in computational programming). Over the course of our interview, we covered bomb simulations, protein model lineage “rivalries”, his work with Karl on AMBER ff15ipq, Moore’s law and the contributions of the hedge-fund billionaire, D.E. Shaw (most notably the supercomputer, Anton). At one point we were on the topic of the obesity epidemic and American healthcare. It seemed to me that Dave, had a lightning-quick mind with a brilliant imagination. This, I imagined, helped him cover ground very quickly, in his coding for the protein models.
Karl was the analyst. His short, precise answer to the question, “Where did you grow up?” reflected the exact and precise way in which his mind worked. I believe this is what made Karl a great coder: the ability to dig deep and systematically write and edit code as an exact and precise solution to the coding problem at hand. I began to see also that coders were not cookie-cutter stereotypes; different minds had different strengths in the realm of coding. If Karl was the coding technician, Dave was the coding artist.
Karl and Dave were developers of AMBER ff15ipq. But many other developers had preceded them, producing AMBER models of which their models were based. So some weeks later, I went to Lillian’s office and asked her who started the AMBER line. She pointed behind me.
On the wall was a flyer for an honorary ceremony of the late Peter Kollman. He was a big man, with big glasses a big beard and an ear-to-ear grin with smiling eyes. He wore his smile with no great effort; you could tell he always smiled.
“He was really larger than life. Actually, he was almost seven feet tall,” Lillian said.
“Wow!” I looked at his picture again and his reassuring smile and kind eyes made me feel bigger too.
“Everyone loved him. When he passed, there was such a collection of people that attended his funeral. He had a big impact on everyone around him,” she said smiling back at the flyer. “He was my advisor and he passed away a year before I had left. And now in the AMBER community, there’s really a feeling that everyone is carrying on his legacy. What he started. They do their work with him in mind.”
“He sounds like a great man. A giant of a man.”
“But do you know what it was that made him great? It was never about him. He cared so deeply about the people around him. He had this strange effect” — and then she looked past me, looking back — “He could walk into a room and make everyone feel just the best about themselves.”
