Author Archives: Joe Laforet Jr.

The End of an Era

And just like that, B-SURF is over. These past six weeks I’ve learned a lot about myself and the things that interest me. I came into this program not knowing what I wanted to do in science, just that I wanted to be a scientist. After spending six weeks hearing about my colleagues’ struggles in the lab moving small amounts of colorless liquid from tube to tube, I’ve decided that wet-lab isn’t for me.
At the beginning of this program the general vibe I got from people was that they felt sorry that my lab work was all virtual. Now, after going into the lab twice to make nanoparticles, I can say that I prefer coding over staring at tubes any day. I also didn’t have to deal with the frustrations of cells/organisms dying on me the day before data was ready to be collected, or the monotony of pipetting samples for hours on end. Throughout this program I was constantly engaged with my work. It was up to me to design the software pipeline for my project. I was given a task by my mentor, and it was on me to implement the features he wanted. One of the challenges that I’ve struggled with most being a self-taught programmer was finding confidence in my coding abilities. I knew that I knew how to code and problem solve, but other than stock problems I had no means to apply my skills. This program was exactly the push that I needed to give me the confidence in myself that I can accomplish problems put in front of me. As the second half of summer, and a condensed semester of organic chemistry, looms ahead of me, I am excited to say that I will continue with my project in the Reker lab. I still need to implement a machine learning model that will hopefully accurately make predictions in nanoparticle formation for me. I’m excited to see what the future holds, and I’m thankful for this experience allowing me to narrow down my search for what I want to do.

Analysis of Hydrogen Bond Formation in Molecular Dynamics Simulations Predicts Formation of Self-Aggregating Nanoparticles

Joe Laforet

Mentors: Zilu Zhang, Dr. Daniel Reker PhD

Department of Biomedical Engineering, Duke University

Co-aggregating nanoparticles can stabilize drugs with more than 90% drug loading capacity. While machine learning can be productively employed to identify nanoparticles, this approach requires large datasets. Simulations provide an opportunity to design nanoparticles without prior data generation, but this method has not yet shown sufficient accuracy. Here, I will develop a novel simulation-based approach that achieves productive accuracy of nanoparticle predictions. By pairing a predictive machine learning model and molecular dynamics simulation software, we analyzed hydrogen bond formation in simulations and used our findings to identify pairs of interest. We compared our predictions against already known data and found that the presence of hydrogen bonding indicates higher likeliness of nanoparticle formation in more than 75% of analyzed pairs. Using this analysis protocol, we plan to analyze and predict other small antiviral nanoparticle formulations aimed at targeting viral diseases such as COVID-19.

Bird Brains and Tech X

This past week we were tasked with going on stage in front of our peers and presenting an eight minute summary of our projects with nothing but a dry erase marker and a whiteboard. While I thought this experience was stressful to prepare for, I am thankful for being able to give an “old-fashioned” pitch of our work. It was interesting to see how much more challenging giving a talk is when your slides aren’t behind you. Personal reflection aside, one talk that captured my interest was George’s presentation on his work in the Mooney lab.

George is doing work with Zebra Finches and a mysterious (for proprietary reasons) biomarking drug called Technology X. Without declassifying anything important about it, Technology X is a drug that was developed with the intent of giving labels to different types of cells in the brain. If it works, it will allow researchers to better study the different parts of the brain on the cellular level. The Mooney lab isn’t responsible for developing this technology, however. They are focused on actually testing it out on live subjects. George’s job this summer is to perform careful neurosurgery on Zebra Finches and deliver Technology X to different parts of the brain.

This project is interesting to me for two reasons. First and foremost, one of my peers is performing neurosurgery on live subjects! Coming from a person who could barely dissect a pig fetus in high school biology, it blows my mind that George is able to work with that degree of precision, and keep his subjects alive after! Secondly, as someone who leans more towards the functionality of things in science, I was really interested in how Technology X works. Sadly I don’t have enough clearance to further probe its mechanisms, but it was still neat to hear about. Overall, this week was enlightening as I was able to see a wide array of the different projects going on in the B-SURF program. I find myself always being boxed into the same functional-type projects, so hearing more about applications and raw research was a welcome change of pace. I’m excited to see what my peers have to present at the poster presentation!

A Day in My Life

What does a virtual lab look like? In short, my days are whatever I make them out to be. Every Wednesday I meet with my mentor Zilu in the new Engineering building and we construct a game plan for the week. Sometimes before our meeting he’ll have me complete a short quiz to familiarize myself with the concepts I’ll use in the coming week. We then go over the quiz together while having a deeper discussion of the concepts covered. This usually takes around half an hour, so for the rest of the day we split off and get to our work for the day. We’re both doing a mix of coding and simulation analysis, so it’s challenging to have an over-the-shoulder mentor relationship. On days where I’m not in lab physically, I either work in the library on East, the library on West, or in my building’s study room. One of the perks of living in the digital age is that Zilu is just a Zoom call away if I need help debugging my code. 

I’d say that I have a love-hate relationship with coding. There’s nothing more satisfying than code working exactly how you intended it to, but that almost never happens. The bulk of what I do is trying to comb through the files I’ve made or scouring forum sites trying to make sense of the error messages I generate. I’m always learning while I work. As this program is going on I find myself running commands without even thinking of them. In the beginning I had to reference my “cheat sheet” for almost every line. Now I can generate .tsv’s and .pdb’s with my eyes closed. Everything I do follows a systematic path. It’s kind of repetitive, but since I’ve done the process hundreds of times now it’s second nature. My days are spent at my keyboard listening to jazz with my fingers dancing away at the command line. Getting in “the zone” is one of my favorite parts of this job. One of my favorite memories so far was when I had a Eureka moment at 3am. The night before I had been struggling with a bug in how to specify the parameters of the simulations. I went to bed grumpy, stewing over the red screen I had been staring at for hours before when inspiration struck. I woke up, grabbed my notebook, and then poured everything out of my head onto the page. After inspiration faded, I went back to bed and then implemented all of the features I had dreamt about in the morning. The most beautiful part? It worked like a dream.

Meet Professor Daniel Reker

Professor Reker has had a combined passion for chemistry and computer science since his high-school days. He was born in and attended school in Germany. The German school system is slightly different from the American school system. In what would be our equivalent of high school, Professor Reker focused primarily on chemistry and English while also minoring in computer science. At this stage in time, due to the lack of truly interdisciplinary, undergraduate educational programs, Dr. Reker was under the impression that he would have to choose between his two passions of computer science and chemistry. Based on the waves that computing was making in the early 2000’s and the resulting promising job market, Reker decided to pursue an undergraduate degree in computer science at the Technical University of Darmstadt, one of the top three computer science programs in Germany.

While he was accepted into a prestigious program, and was enjoying the computer science curriculum, Reker still felt like he was missing something in his education. Computer science appeared to be a powerful tool that should be applied to other domains to positively impact people’s lives. He was still passionate about biology and chemistry, so in order to stay with these fields he became involved with multiple extracurriculars in the biology department, leaning towards computational-flavored topics, and regularly attended seminars. After undergrad, Dr. Reker graduated with a degree in computer science, with added experience in computational biology. From here, he decided to further explore the overlap between computing and biology and earned a Master of Science from ETH Zurich in computational biology and bioinformatics.

As a master’s student, Professor Reker was especially drawn to the pharmaceutical science department of tETH Zurich and was interested in the application of algorithms for analysis and development of new drug candidates. After his masters, Reker stayed with the pharmaceutical science department and completed a Doctor of Science (equivalent to a PhD) with Dr. Gisbert Schneider in the “Computer-Assisted Drug Design” laboratory. After completing his PhD, Reker was yearning for more than just theoretical development of tools. He wanted to have a more translational impact, and actually use the tools he develops towards improving therapeutic options for patients. This drove him to the lab of Dr. Robert Langer at MIT, where Reker completed a post-doc in drug delivery, a later step in the drug development process with thereby higher chances of translation.

Staying true to his interdisciplinary roots, Dr. Reker’s vision for his lab is one that bridges multiple disciplines. As the lab gets more established in the new Wilkinson building, he aims to have roughly a 50/50 split between computational work and wet lab research with every student being involved in both aspects. The goal is to have a team made of people from a broad expanse of backgrounds. Having all of these different flavors of science working on the same projects greatly enriches the educational enterprise and the science, as there are multiple approaches being taken towards solving challenges in drug discovery and delivery. Professor Reker’s vision is one in which the lab is, “…an incubator space where students can learn to become multilingual in the different aspects of lab work and to come to learn what the hype around machine learning for pharmacology is about.”

Mentorship is also one of the most important aspects of science to Professor Reker. Even on the first day of lab meetings, I felt welcomed as Dr. Reker emphasized that his goal was to mentor students to hopefully surpass him as a scientist, rather than himself being a master of everything. This attitude has been something that has been with Dr. Reker for most of his academic career but was ultimately cemented as one of his core beliefs after his experience being an undergraduate TA for a very difficult computer science course. He ended up mentoring a group of students and the experience was so profound for them that they decided to become TA’s after him.

Wrapping up this interview, I learned a lot about Dr. Reker, and the journey involved with being an interdisciplinarian. Like Dr. Reker, I also find myself called to computer science, but still needing chemistry and biology in my life. Under Professor Reker’s mentorship, I hope to follow in his footsteps at the intersection between these three branches of science and become a “multilingual” student.



Novel Nanoparticles

If you were to ask me a week ago what a nanoparticle was, I’d give a pretty generic answer saying that it was something really small. While that statement is true, there is so much more to these microscopic specks than we understand. One of the biggest problems in drug development is the issue that not all compounds are soluble in water. Not being able to be dissolved means it’s hard for the body to absorb the medicine that you are trying to take. There is also the issue of directing the drug to its intended target. A potent liver drug isn’t useful if it accidentally makes its way into your heart. These two issues can be solved with nanoparticles.

Research into the design of different types of nanoparticles has shown promise, but often results in a combination of a difficult manufacturing process and a low, around 5-10%, drug-loading capacity. The Reker lab is taking a novel approach to form a different kind of nanoparticle. Our nanoparticles rely on the natural tendency of select molecules to aggregate together and form clumps. These “clumps”, really nano-sized clusters, of drug and excipient pairs have a revolutionary potential for up to 95% drug-loading capacity. They are also significantly easier to manufacture than traditional nanoparticles. My project this Summer is building off of Dr. Reker’s previous work in order to find a potential treatment for Covid.

Imagine a conveyor belt in a car factory. All of the different parts for the car go along the belt, and are assembled into larger parts at each stop. My day-to-day involves working at one of these stops. Currently, my “stop” on the conveyor belt deals with simulating the potential interactions of drug-excipient pairs of interest. I get a pair found from a machine learning model, run some code, generate a simulation, and then analyze the simulation and see if it shows promise for a nanoparticle to form. It’s a pretty straightforward process, until you realize that the simulations can take hours to generate. Simulations are also an approximation of what could happen in reality, not necessarily what will happen. This approximation can be made better by fine-tuning the generation parameters. Part of my job is to find the “sweet spot” combinations that most accurately reflect what would happen if we were to make the particles in the lab. Eventually, if I find a combination that shows promise, I will get the opportunity to try synthesizing my own nanoparticles.

While most of my research has been dry work, I’m starting to find analyzing the simulations fun. Seeing all of the cool movies from visual representation is satisfying. I also do some analysis on intermolecular interactions, but that’s more numbers-based. My experience so far has been filled with all of the things that I like about science. There is so much overlap with this project in particular. I need to understand how the code runs in order to generate simulations, I need to know how to code in order to analyze the intermolecular interactions, I need to have a good understanding of chemistry to make sense of the simulation, and I need to have an open mind as it’s only me and my mentor working with this pipeline. We are the pioneers in this space. It is up to us to find the optimum parameters and develop analysis techniques ourselves. It’s both intimidating and exciting being alone in this space, but I’m loving the journey so far and I can’t wait for what the future holds!

This is one of the cooler simulations I’ve run. Normally, the particles form clumps with each other, but this one has excipient molecules on vertices of the drug cluster forming a starfish shape.


Diving into the Realm of Quantum Simulation

     Nanoparticles, quantum mechanics, molecular dynamics, supercomputers, and machine learning. All five of these topics sound like things that Tony Stark and Bruce Banner work on in their free time. However, with modern advances in computing technology these fields are actually able to exist and coexist in a beautiful way giving rise to the real-life field of nanomedicine. 

   I’m coming into the Reker lab as a rising sophomore BME student still trying to figure out what I actually want to study. My interests are spread over multiple disciplines, and it’s hard to pin down exactly one thing that I like. Growing up I loved playing with computers. Not building them like Jimmy Neutron, but simply playing video games like Minecraft. I loved how the computer was able to generate an environment that had similar rules to reality. You have objects in the world, collision rules, like when you jump on slime blocks, gravity, and a day-night cycle. Each of these parts of the simulation combine in a unique way to create the “world” and give a fun experience while playing the game. The combining of multiple processes to generate a real-life simulation is basically what Dr. Reker and other computational chemists are doing with Molecular Dynamics.        

 Before the B-SURF program I had no idea that computational chemistry even existed. After reading through some papers and meeting with Dr. Reker and my mentor Zilu, I’ve learned that computational chemistry could revolutionize the field of drug development. Through simulating interactions of molecules with the Molecular Dynamics software, we’re able to predict what the resulting body would be if we combined the reactants in real life. Specifically, I’m going to spend my Summer researching the potential formation of nanoparticle complexes between different drug and excipient molecules. While most of the chemistry is flying a bit over my head, I hope that as the Summer goes on, and as I finally get around to taking Organic Chemistry, I will understand more about the chemical details of why and how our simulations of molecules form. I’m also excited to learn more about optimizing the parameters of the simulation to perform more accurate predictions. For now though, I’m happy with this past week’s work and my little knowledge in how to generate simple simulations. I might just be a big nerd, but I find it magical that we’re able to take a simple string of letters and pass through a coding pipeline and out comes a visual model of molecules interacting with each other.

     While I don’t get to go to an in-person lab like my fellow B-SURFer’s, I wouldn’t want to change anything about my experience so far. I’m a learn-by-doing kind of guy, and the flexibility of working in a computational lab fits that perfectly. I also know that as time goes on, and as Dr. Reker’s lab gets established in the new engineering building, we’ll eventually get to translate our simulations into real-life. Overall, I’m very excited for what the future holds, and am extremely grateful for Dr. Reker and Zilu for giving me the chance to study underneath them. Here’s to the start of a great summer!

Here is one of my first simulations. In this example, I simulated the interactions between the anti-cancer drug Sorafenib and Cholic Acid. I’m still working on finding the best way to make a movie to upload, so stay tuned for more cool gifs.