Mining the National Cancer Institute's Tumor-Screening Database:  Identification of Compounds with Similar Cellular Activities

Abstract

In an effort to enhance access to information available in the National Cancer Institute's (NCI) anticancer drug-screening database, a new suite of Internet accessible (http://spheroid. ncifcrf.gov) computational tools has been assembled for self-organizing map-based (SOM) cluster analysis and data visualization. A range of analysis questions were initially addressed to evaluate improvements in SOM cluster quality based on the data-conditioning procedures of Z-score normalization, capping, and treatment of missing data as well as completeness of drug cell-screening data. These studies established a foundation for SOM cluster analysis of the complete set of NCI's publicly available antitumor drug-screening data. This analysis identified relationships between chemotypes of screened agents and their effect on four major classes of cellular activities: mitosis, nucleic acid synthesis, membrane transport and integrity, and phosphatase- and kinase-mediated cell cycle regulation. Validations of these cellular activities, obtained from literature sources, found (i) strong evidence supporting within cluster memberships and shared cellular activity, (ii) indications of compound selectivity between various types of cellular activity, and (iii) strengths and weaknesses of the NCI's antitumor drug screen data for assigning compounds to these classes of cellular activity. Subsequent analyses of averaged responses within these tumor panel types find a strong dependence on chemotype for coherence among cellular response patterns. The advantages of a global analysis of the complete screening data set are discussed.

Linking pathway gene expressions to the growth inhibition response from the National Cancer Institute's anticancer screen and drug mechanism of action

“…We have previously established the general MOA of the agents in some of these SOM regions: mitosis (M), membrane function and oxidative stress (N), nucleic acid metabolism (S), and metabolic stress and cell survival (Q), oxidative metabolism (R), and kinases/phosphatases and oxidative stress (P), via other methods. 16,23,24 The pathway mapping results provide additional support for the annotation of some of the SOM regions: for example, the GO terms mitotic checkpoint, cytokinesis, kinetochore, and cell cycle are associated with the M-region; the GO terms mitochondrial inner membrane (data not shown), response to oxidative stress (data not shown), and oxidoreductase activity are associated with the N-region; the KEGG pathway one carbon pool by folate, the BioCarta granzyme Amediated apoptosis pathway, the GO terms DNA topological change and DNA topoisomerase (ATP-hydrolyzing) activity, are associated with the S-region; the KEGG pathway glutamate metabolism and the GO terms (data not shown) xenobiotic metabolism, cysteine metabolism, and glutathione biosynthesis are associated with the Q-region; the KEGG pathways fatty acid metabolism and oxidative For each pathway, the table lists the name of the pathway, the significance level of the connection between the pathway and its small-molecule inhibitors (P-value), and the number of references where the small molecule is documented as an inhibitor of the product of a gene in the pathway (see Supplementary Information, Table IV, Galactose metabolism N hsa00252…”

“…21 GI 50 growth patterns have been found to be an informationrich resource for establishing a compound's MOA. 16,[22][23][24] SOM clustering of the GI 50 data segregates compounds into nine major response categories: mitosis (M), membrane function (N), nucleic acid metabolism (S), metabolic stress and cell survival (Q), kinases/phosphatases and oxidative stress (P), and four unexplored regions R, F, J, and V. 16,23,24 Each of these regions is further divided into a total of 80 clades (a clade is a group of clusters (nodes) that share similar cytotoxic responses) (subregions: M 1 -M 8 , N 1 -N 13 , P 1 -P 8 , Q 1 -Q 7 , R 1 -R 7 , S 1 -S 13 , F 1 -F 8 , J 1 -J 8 , V 1 -V 8 ) (Figure 1). The current SOM extends our previously published analysis to include the existing complement of newly screened compounds.…”

“…The current SOM extends our previously published analysis to include the existing complement of newly screened compounds. 23 Gene expression patterns across the NCI 60 can be organized in terms of predefined pathways or functional Figure 1 Cytotoxicity measurements for nearly 30 000 small molecules screened against the NCI's tumor cell panel (NCI 60 ) were clustered into 1350 groups or nodes using an SOM. 23 Each cluster is represented by a hexagon and neighboring clusters identify compounds with similar cytotoxicity responses.…”

“…23 Gene expression patterns across the NCI 60 can be organized in terms of predefined pathways or functional Figure 1 Cytotoxicity measurements for nearly 30 000 small molecules screened against the NCI's tumor cell panel (NCI 60 ) were clustered into 1350 groups or nodes using an SOM. 23 Each cluster is represented by a hexagon and neighboring clusters identify compounds with similar cytotoxicity responses. These nodes are separated into nine major response categories (regions): mitosis (M), membrane function (N), nucleic acid metabolism (S), metabolic stress and cell survival (Q), kinases/phosphatases and oxidative stress (P), and four unexplored regions R, F, J, and V. Each of these regions is colored differently and further divided into a total of 80 clades (a clade is a group of clusters that share similar cytotoxic responses) (subregions: M 1 -M 8 , N 1 -N 13 , P 1 -P 8 , Q 1 -Q 7 , R 1 -R 7 , S 1 -S 13 , F 1 ÀF 8 , J 1 -J 8 , V 1 -V 8 ), shown here by the black boundary lines.…”

Chemical Data Mining of the NCI Human Tumor Cell Line Database

“…In their 2000 paper, Scherf et al 7 examine correlations between compounds' high-throughput screening results (the activity pattern set) and mRNA expression levels. Recently, Rabow et al 8 performed a clustering of the tumor cell line data set based on the activity profiles, using a self-organizing map (SOM). Other work at the NCI focused on ellipticine analogs and the potential relationship between the mechanism of action and the 60 cell line activity profiles.…”