Faris Fizal, a high school student who has been working with us since January, was interviewed by the CBC. Faris is preparing for the Sanofi Biogenius Canada competition, where his project involves molecular simulations of natural and engineered variants of hemoglobin to be understand their stability under shear stress, with the ultimate goal to produce synthetic blood replacements.
Our new cluster, GLaDOS, is coming together.
The system features:
- 96 GTX 1080 Ti GPUs
- 208 CPU cores
- 3072 GB of RAM
- More than 1.1 PB of raw storage space
We decided to do a team building exercise at one of the escape rooms at The Locked Room. Someone with a sense of humour thought it would be fun for us to try the hardest room. We did not escape, but it was a ton of fun.
We just posted a new pre-print to bioRxiv on the optimization of Hamiltonian replica exchange simulations. The abstract is below:
Replica exchange is a widely used sampling strategy in molecular simulation. While a variety of methods exist for optimizing temperature replica exchange, less is known about how to optimize more general Hamiltonian replica exchange simulations. We present an algorithm for the on-line optimization of both temperature and Hamiltonian replica exchange simulations that draws on techniques from the optimization of deep neural networks in machine learning. We optimize a heuristic-based objective function capturing the efficiency of replica exchange. Our approach is general, and has several desirable properties, including: (1) it makes few assumptions about the system of interest; (2) optimization occurs on-line wihout the requirement of pre-simulation; and (3) it readily generalizes to systems where there are multiple control parameters per replica. We explore some general properties of the algorithm on a simple harmonic oscillator system, and demonstrate its effectiveness on a more complex data-guided protein folding simulation.
Feedback is welcome.
We just posted a new pre-print to bioRxiv. on the role of physical modeling in the future of structural biology. The abstract is below:
Heuristics based on physical insight have always been an important part of structure determination. However, recent efforts to model conformational ensembles and to make sense of sparse, ambiguous, and noisy data have revealed the value of detailed, quantitative physical models in structure determination. We review these two key challenges, describe different approaches to physical modeling in structure determination, and illustrate several successes and emerging technologies enabled by physical modeling.
Feedback is welcome.
It was a big week in the lab, with three papers submitted. Target journals are Angewandte Chemie, Current Opinion in Structural Biology, and Journal of Physical Chemistry B.
I just returned from the 3rd International Meeting on Protein and RNA Structure Prediction held in Montego Bay, Jamaica. It was an excellent meeting held in a beautiful location.
I presented our work on work on determining structures from sparse NMR data, in collaboration with the groups of Guy Montelione and Chris Jaroniec. There was a lot of interest and I had several interesting discussions.
There were a lot of interesting talks on topics ranging from better ways to do SAXS experiments to new structure prediction algorithms.
There was also some free time to explore. We went to a luminous lagoon filled with bioluminescent plankton that emit light in response to shear stress. It was an amazing experience.
Overall, it was a fantastic conference that I look forward to returning to in 2 years time.
Emiliano Brini presents his work on using MELD to predict protein structures and interactions.
Our new paper "Molecular modeling of biomolecules by paramagnetic NMR and computational hybrid methods" is now available from BBA Proteins and Proteomics.
This is intended as an easy to digest review of paramagnetic NMR methods and associated computational modelling strategies.
I'm extremely proud to announce that the MacCallum lab was awarded a five year grant from CIHR for our project aimed at developing new molecular diagnostics for crystal arthropathies. This is an exciting project with great results to date, and we're grateful for funding to push it further.
Overall success rates in this year's CIHR Project Grant were only 16 percent, which is completely unhealthy and reflects funding that has been flat for 15 years. Success rates outside of BC and Ontario were half the national average. Such regional disparities will not lead to the thriving research environment that Canadians need.
While I'm fortunate to have obtained a grant in the in the current climate, I hope that the Minister of Science recognizes the need for improved funding for science in Canada.