Critical Renewal of an Obsolete Computational Biochemistry Suite at the University of Lethbridge

莱斯布里奇大学的Wetsel和Wetmore实验室是使用计算和数学方法研究生化系统的领导者。我们团队研究的问题在空间尺度上从分子到器官水平，在时间尺度上从飞秒到几周。所使用的技术来自计算化学，非线性动力学和随机动力学。研究的问题包括污染物的致突变性，核酸-蛋白质相互作用，生化网络结构在细胞命运决定中的作用，以及基因表达延迟在生化动力学中的影响。具有国际竞争力的计算生物化学实验室需要高端计算机来有效地进行模拟，分析和可视化，以及为加拿大计算/计算（CCC）高性能系统准备工作。尖端的台式计算机最大限度地提高了吞吐量，减轻了CCC设施的压力，这些设施可以用于最苛刻的工作。特别是，较小的计算以及较大作业的测试和优化可以在本地进行，避免了CCC CPU，特别是GPU周期的浪费。我们的计算生物化学设施，位于国家的最先进的科学公地复杂，旨在最大限度地提高实习生主管和实习生实习生的互动，与教师办公室直接毗邻我们的实验室，加速研究，并导致实验室在白天的高占用率。由于学生每天都在实验室工作，台式计算机不能共享。计算科学的公平性需要随时获得高端计算，所要求的设备将确保这一点。 不同的高端工作站配置需要执行的各种任务。例如，AMBER（分子动力学模拟）在GPU上的运行速度比在只有CPU的机器上快约500倍; Python中的图形算法几乎与CPU核心数量成线性关系;这两者以及Maple符号代数计算（CCC硬件上不可用）通常需要大量内存空间。目前，Wetmore和Wetsel实验室拥有不同年份的计算机，其中大约一半是8-10岁。硬件故障和过时，这阻碍了公平地获得更新的软件，正在侵蚀我们的能力，以推进我们的研究计划和提供学生培训。要保持我们实验室的世界级竞争力和培训能力，就需要紧急更换更多的计算机，而这些计算机实际上是用运营赠款购买不到的。因此，要求提供资金，以具有高端能力的现代化工作站取代我们计算实验室27个工作站中的14个。由于目前的本地存储不足以处理我们现在常规生成的较大数据集，我们正在请求一个补充磁盘阵列，以实现后处理和可视化，同时最大限度地减少CCC阵列之间的耗时传输。

The Roussel and Wetmore labs at the University of Lethbridge are leaders in studying biochemical systems using computational and mathematical methods. The problems our teams study range in spatial scale from the molecular to the organ level, and in time scale from femtoseconds to weeks. Techniques used derive from computational chemistry, nonlinear dynamics and stochastic kinetics. Problems studied include the mutagenicity of pollutants, nucleic acid-protein interactions, the role of biochemical network structure in cell fate determination, and the effects of gene expression delays in biochemical dynamics. Internationally competitive computational biochemistry labs require high-end computers to efficiently conduct simulations, analyses and visualization, as well as preparation of jobs for Compute/Calcul Canada (CCC) high-performance systems. Cutting-edge desktop computers maximize throughput and alleviate the pressure on CCC facilities, which can then be reserved for the most demanding jobs. In particular, smaller calculations as well as testing and optimization of larger jobs can be carried out locally, avoiding waste of CCC CPU and, especially, GPU cycles. Our computational biochemistry facilities, located in the state-of-the-art Science Commons complex, are designed to maximize trainee-supervisor and trainee-trainee interactions, with faculty offices directly adjacent to our labs, accelerating research and resulting in high occupancy of the labs during the day. Desk computers cannot be shared given that students are working in the lab every day. Equity in computational science requires ready access to high-end computing, which the requested equipment will ensure.  Different high-end workstation configurations are needed for the variety of tasks carried out. For instance, AMBER (molecular dynamics simulations) runs about 500 times faster on GPUs than on CPU-only machines; graph algorithms in Python scale almost linearly with the number of CPU cores; and both of these as well as Maple symbolic algebra computations (not available on CCC hardware) often require large memory spaces. The Roussel and Wetmore labs currently house computers of different vintages, roughly half of which are 8-10 years old. Both hardware failures and obsolescence, which hinders equitable access to updated software, are eroding our ability to advance our research programs and provide student training. Maintaining world-class competitiveness and training capacity in our labs requires the urgent replacement of more computers than can realistically be purchased with operating grants. Funding is therefore requested to replace 14 of 27 workstations in our computational labs by modern workstations with high-end capabilities. Since current local storage is inadequate to handle the larger data sets we now routinely generate, we are requesting a complementary disk array to enable post-processing and visualization while minimizing time-consuming transfers to and from CCC arrays.