全部 标题 作者
关键词 摘要

OALib Journal期刊
ISSN: 2333-9721
费用:99美元

查看量下载量

相关文章

更多...

CheS-Mapper - Chemical Space Mapping and Visualization in 3D

DOI: 10.1186/1758-2946-4-7

Full-Text   Cite this paper   Add to My Lib

Abstract:

The area of quantitative structure-activity relationship (QSAR) modeling, an active field in chemoinformatics, includes numerous approaches, many based on numerical features, structural fingerprints, or sub-graph mining [1-3]. Commonly, the employed features are then used with machine learning approaches like support-vector-machines or decision trees to predict activities for previously unseen compounds. However, many algorithms do not convey information about the relevance of features with respect to particular activity values in the learnt models. Visualization can be vital to examine the dataset for possible interdependencies between compound structures, compound features, and endpoint values. In the work described here, we present a general, interactive, and open-source application called CheS-Mapper, for the inspection of chemical datasets of small molecules, enabling scientific researchers to experiment with the compounds and their features. Our application visualizes compound structures and features in a virtual 3D space. The tool detects subgroups (clusters) within the data, and can be employed to analyze the data to find possible structure-activity relationship (SAR) information.We give a short overview of existing visualization methods. The probably most commonly used tool to explore a small molecule dataset is the chemical spreadsheet. It lists all compounds, usually with a small picture showing its 2D structure, and other properties in one single table. There are several commercially available spreadsheet programs [4-6]. The open-source workframe Bioclipse [7] has also a built in spreadsheet tool. These tools often incorporate plug-ins to show the 3D structure of single compounds, and plotting functions like heat-maps for compound properties. A common plotting method, that is frequently available within chemical spreadsheet programs as well, is based on multi-dimensional scaling [8]. This technique reduces the high-dimensional input feature space in orde

Full-Text

comments powered by Disqus

Contact Us

service@oalib.com

QQ:3279437679

WhatsApp +8615387084133

WeChat 1538708413