This week, I presented my work to the whole lab group, and I finished writing and editing my final report. I’m really going to miss all the incredible people that I’ve met and the friends that I’ve made here in Houston.

In my final week of work, there were a couple things I was finishing up. Firstly, I made a small dataset of cleaned-up PDB files for Class II MHCs – the data for these structures is pretty inconsistent, so it’s good to have a set of clean, usable files. I’ve also been working more with SLURM and the HPC cluster. The grad student I’ve been working with, Romanos, is planning to compare the results of this tool, APE-Gen2.0, to other pMHC modelling tools out there. To do this, he’s running all of the them on the PANDORA dataset, and since he didnt have the results from APE-Gen1.0 on this dataset, I began running it. It’s cool to actually understand how the Cluster works and be able to run jobs on it – I’ve never had that chance before!

This week, I finally got some preliminary results for modelling Class II MHCs – and they’re very promising. The orignal concern was that the workflow that’s currently being used for Class I MHCs wouldn’t be transferable to Class II MHCs because the two classes hold their peptides differently. Class I MHCs deal with shorter peptides, have a deeper binding cleft, and the peptide forms an arch; Class II MHCs have longer peptides that often have free-floating ends, the binding site is between two protein chains (rather than one) and the peptide is held much straighter. Random Coordinate Descent (RCD) is the approach used to come up with different peptide configurations, and since it works really well on the Class I arch-shaped peptide, there was some concern that it wouldn’t work as well on the straighter Class II peptide. However, I gather some examples, ran RCD, and determined that the current workflow should work!

My friends and I also visited the NASA Space Center – it was super cool to wander around the exhibits, see the Apollo Mission Control Center, and grab some famous Texas fried chicken on the way back.

This week, I was assigned my own chunk of the project – looking into modelling Class II MHCs. In structure, these complexes are different from the Class I MHCs because the binding side is formed from two amino acid chains, rather than just one. In function, Class II MHCs present exogenous (foreign) peptides to immune cells, so they’re typically associated with pathogens like viruses and bacteria, whereas Class I MHCs present endogenous/cytoplasmic peptides, which can be indicators for cancer and cell health monitoring.

Now that I’m about halfway through my tme here at Rice, I’ve started to wrap up some of the changes I’ve been working on and start switching gears. This was a short week for me, given the holiday and some travel delays, but I was able to look into how flexible residues of the receptor are fixed, which is framed as a constraint satifaction problem. I spent a really long time looking through PDB files to understand the changes that are being made and how the current code is working, and while I feel like I have a better sense of how it’s working, I’ve been struggling to find a way to improve the efficiency. I hope that within the next few days I’ll be able to find something promising.

This week, I focused on merging in all new changes I’ve been making, as well as studying the constraint satifaction problem that’s being solved to determine the labels of atoms of the peptide’s amino acids. Merging in the changes took a lot longer than I had anticipated, but I’m glad I put in the time to test and make sure I wasn’t breaking anything. This CSP isn’t related to anything I’ve worked with before, so I ended up going through the code line-by-line to mkae sure I understand what’s happening and why. I’ve also been reading a few more papers and getting a better sense of the current work being done in this area.

This week, I continued to work on some improvements to the APE-Gen codebase and got to know my labmates a bit better. Though everything is still pretty new to me, I feel as though I’m starting to really understand how all the different parts of the code are linked together and I’m finally grasping how all the packages work. Increasingly, I’m realizing how valuable my computer science education, as well as my experiences as a TA, since I’m able to understand more complex pieces of code easily and write really clean, modular code as well. I’ve really enjoyed lab meetings – both with the smaller bio group as well as full-lab meetings – as they’ve given me a better sense of the graduate student community and lifestyle in addition to being helpful and thought-provoking. I’ve gotten a chance to do some of the tourist activities around Houston; over the weekend I went to Galveston beach, which was so hot but very fun.

Unfortunately, I went into this week with a slight cold, so I was a bit out of it on Monday and Tuesday, but it was amusing that, given this project’s application in the immune system, I was thinking about the immune responses my body was putting up and how this project can model certain parts of that response. Once I was feeling better, I finalized and fixed up the formatting for the changes I made from the previous week, then I started on fixing up some of the functions throughout the program. Mainly, I followed the improvements listed on the project’s README, but as I’ve become more familiar with the code, I’ve found a lot of areas that could be cleaned up and made more versatile. With the blessing of the PhD student I’ve been working alongside, I’ve been considering ways that I could modularize the tools that this program already implements well, and made it more accessible to biologist who may not have a strong computer science background. I’ve found that I’ve been able to settle into my role here smoothly, and I’ve really enjoyed meeting the other members of the lab as they come in-person, and it’s been super nice to have other undergrads to work near during the day.

At the start of this week, I met with the graduate student I’ve been working alongside, Romanos. Romanos had been so incredibly helpful, walking me through more complicated steps in the code and answering my barrages of questions with so much patience and enthusiasm. Now that I have a good sense for how the code works and how the packages work together, I’ve begun working on the tasks that have been assigned to me. I started with some of the easier tasks, such as renaming the intermediate files that are generated as the program runs and shifting some repeated code to be a macro. I’m a bit surprised by how much I’ve already learned and how quickly I’ve been able to contribute to the project, even if it is minor – I was bracing for a much longer on-boarding experience, but I’m really glad I’m able to be productive so soon into this summer. Some other things I changed included fixing a persistent warning, expediting the startup process, and starting on some of the class documentation. Beyond work, I’ve been getting to know the area a bit better, but the temperature and humidity are a bit higher than I’m used to.

This week was my first week in Dr. Kavaraki’s lab. It was a pleasure to meet the lab manager, Dr. Kingston, as well as the members of the biology group. At my first meeting, I got a better sense for the grad student’s projects, including which project I’ll be working on. This project, headed by Romanos, involves modeling and predicting how a given peptide will bind to a Major Histocompatibility Complex (MHC), which is a protein complex core to the immune system. I spent the rest of the week digging through the code for this project, learning the ins and outs of the packages and processes being used. I had an excellent time getting to know the members of the lab, attending meetings, and exploring Rice’s campus. While Texas in the summer is a bit warmer than I’m used to, it seems like a fun city and I’m excited to spend the next few months here!