|
TreeThreader: Thousands of volunteers build a web server for protein structure prediction
We are glad to announce that the new application TreeThreader is officially launched on CAS@home. TreeThreader is the second generation software developed by scientists from ICT (Institute of Computer Technology, Chinese Academy of Sciences) and MU (Digital Biology Laboratory, Missouri University) for protein structure prediction and it outperforms the first generation software SCThreader by 2 times in speed together with significant improvement in accuracy.
ICT-MU scientists also provide a web service
(http://protein.ict.ac.cn/TreeThreader) for biologists to submit their protein sequences. The structure predictions of these proteins are
performed on the thousands of volunteer computers via CAS@home platform, which makes it possible to perform proteomic-wide prediction, say, prediction structures for all proteins of human stomach cancer cells.
Protein structure prediction in an computing-intensive task. For instance, threading, the leading methods for protein structure
prediction, is exceedingly time-consuming because the query sequence should be aligned against a lot of known protein structures. Volunteer computing is absolutely a great platform for protein structure prediction due to the intrinsic high parallelism on thousands of computers.
At this moment we have about 8000 protein sequences to predict which requires about 36.5 years of CPU time on a single core machine. We
strongly invite you to participate in TreeThreader project to contribute to biology and life science.
Yours sincere,
Dongbo Bu , Jin Li (ICT scientists)
-------------------------------------------------
TreeThreader: 千万志愿者构建的蛋白质结构预测网络服务
我们很高兴通知大家,新的应用TreeThreader即将在CAS@home志愿平台上正式发布。TreeThreader是计算所和密苏里大学的科学家们共同开发的蛋白质结构预测算法,是继SCThreader的第二代预测算法,其在速度上将超SCThreader高达2倍,并且在 准确度上也有显著的提高。
ICT-MU的科学家基于这个新的算法共同开发了TreeThreader网络服务 (http://protein.ict.ac.cn/TreeThreader/),为广大的生物学家提供了便捷的 预测服务。在CAS@home平台上,千万志愿者参与这些大量的预测,使蛋白质组全预测成为可能,比如预测人类所有胃癌细胞的蛋白质的结构。
蛋白质结构预测是一项计算密集型的任务。比如,蛋白质结构预测的主流方法——Threading,是极其耗时的,这是因为每个蛋白质序列必须要和 大量的已知结构的蛋白质进行比对。而志愿计算平台因为千万台计算机之间内在的高并行性,决定了志愿计算平台非常适合于蛋白质结构预测应用。
我们目前具有大约8000条蛋白质序列需要预测结构,如果在单机上运行,这将耗费大约36.5年的CPU时间。我们真诚地邀请您参与 TreeThreader项目,为生物学,乃至生命科学做出您宝贵的贡献。
谢谢您的支持!
卜东波, 李瑾 (中科院计算所)
To opt out of future emails from CAS@home, please edit your project preferences at http://casathome.ihep.ac.cn/prefs.php?subset=project |
评分
-
查看全部评分
|