本帖最后由 vmzy 于 2014-1-9 11:27 编辑
An Introduction to Markov State Models
January 8, 2014 by Vijay Pande ·
 For researchers curious about how FAH efficiently uses thousands to hundreds of thousands of processors, we’ve recently put together a new book which describes these Markov State Model methods in detail.The aim of this book is to explain the importance of Markov state models to molecular simulation, how they work, and how they can be applied to a range of problems. The Markov state model (MSM) approach aims to address two key challenges of molecular simulation: 1) How to reach long timescales using short simulations of detailed molecular models 2) How to systematically gain insight from the resulting sea of data MSMs do this by providing a compact representation of the vast conformational space available to biomolecules by decomposing it into states—sets of rapidly interconverting conformations—and the rates of transitioning between states. This kinetic definition allows one to easily vary the temporal and spatial resolution of an MSM from high-resolution models capable of quantitative agreement with (or prediction of) experiment to low-resolution models that facilitate understanding. Additionally, MSMs facilitate the calculation of quantities that are difficult to obtain from more direct MD analyses, such as the ensemble of transition pathways. This book introduces the mathematical foundations of Markov models, how they can be used to analyze simulations and drive efficient simulations, and some of the insights these models have yielded in a variety of applications of molecular simulation.
大意： 最近我们出了本新书，详细介绍了马尔可夫状态模型法，阐述了FAH是如何高效利用成千上万处理器资源的。 本书的目的是解释了马尔可夫状态模型法在分子模拟方面的重要作用。它解决了2个关键问题： 1、如何利用短时间模拟的分子模型细节，完成长时间模拟。 2、如何从海量的数据中系统挖掘有用的信息。 MSM通过把大量的分子生物分构型信息压缩为不同的状态以及状态间的转换速度信息。这种动力学描述，可以轻易的将高分辨率的海量信息在时间与空间上压缩为低分辨率的易理解的信息。此外，MSMs还可以推算出传统分子动力学算法无法直接计算出的信息。 本书介绍了马尔可夫状态模型法的数学原理，以及如何利用来高效的进行模拟。 有兴趣的可以看下FAH团队的讨论会在线视频（译注：需翻墙，你懂的）： http://www.youtube.com/embed/0pB3pUXULmo?feature=oembed
Webinar: Learn about OpenMM, the code driving FAH GPUs
January 8, 2014 by Vijay Pande ·
For those that are curious about OpenMM, the code that drives the GPU calculations in FAH, please join us on January 16th for a webinar presented by Professor Vijay Pande. I’ll talk about about OpenMM, the computational engine behind the Folding@Home distributed computing system, and hear how others have leveraged its high-performance on GPUs and its custom classes and extreme flexibility to accelerate research in areas such as free energy calculations, protein folding, and protein conformational change.
This webinar is planned for January 16th 2014 at 9.00 AM Pacific Time. Space is limited, so please register at http://bit.ly/OpenMM
大意：
将于1月16日，Vijay Pande讲授召开有关GPU核心算法OpenMM的网络研讨会。名额有限，有识之士尽快去注册报名。 | 
发表于 2013-12-31 10:27:40
