2. Query results for DrugBank-based in the pipeline

2.1. In the example of AHF for synonym input (DrugBank)

2.1.1. The query results for drug information

When users input a drug related keyword, such as “AHF” and then click the “Query” button, the waiting page for “DrugBank Query” will be shown as Fig. 2.

Fig. 2. The waiting page for “DrugBank Query”.

After waiting for a moment, the query results for drug information will be shown as Fig. 3. The drug information includes DrugBank [1] ID (e.g., DB00025; arrow line 1 in Fig. 3 for hyperlinking to DrugBank in Fig. 4), common name, chemical formula, molecular weight, description for the DrugBank ID, and DrugBank information linkage. Users can select the desired

(e.g., arrow line 2 in Fig. 3), go to waiting page (Fig. 5) and then pass for getting further DrugBank information (Fig. 6).

Fig. 3. The query results for drug information which the input drug related key is “AHF”.

Fig. 5. The waiting page for “DrugBank Query”.

2.1.2. The query results for the drug targets, metabolizing enzymes, drug transporters, and drug carriers

After waiting for a moment, the query results for the drug targets (arrow line 1) and metabolizing enzymes (arrow line 2) will be shown in Fig. 6. There is no available record for drug transporters and drug carriers. The results of clicking arrow line 1 and arrow line 2 are shown in Fig. 8 and Fig. 9, respectively.

Fig. 6. The query results for the information of drug targets, metabolizing enzymes, drug transporters, and drug carriers to the DrugBank ID “DB00025”.

2.1.2.1. Drug targets (Clicking arrow line 1 in Fig. 6)

The results show 11 drug targets, i.e., (1) Coagulation factor X, (2) Coagulation factor IX, (3) von Willebrand factor, (4) Phytanoyl-CoA dioxygenase, peroxisomal, (5) Asialoglycoprotein receptor 2, (6) 78 kDa glucose-regulated protein, (7) Calreticulin, (8) Calnexin, (9) ERGIC-53 protein, (10) Low-density lipoprotein receptor-related protein 1, and (11) Multiple coagulation factor deficiency protein 2.

Users can click the “Coagulation factor X” (arrow line 1 in Fig. 6) to get further STRING [2] protein interaction information. When users click the “Coagulation factor X” button, the waiting page for “STRING protein Query” will be shown as Fig. 7.

Fig. 7. The waiting page for “STRING protein Query”.

After waiting for a moment, the query results for interested drug target “Coagulation factor X” which have “Preferred Name” and “Annotation” will be shown as Fig. 8. In this case, six interested genes such as “F9”, “F10”, “F8”, “F7”, “TFPI”, and “F3” contain in the output. On the top of this page, the selection for tagSNPs of HapMap parameters, such as HapMap database version [3], population (CEU, CHB, JPT, and YRI), pairwise methods, R square cut off, minor allele frequency (MAF) cut off, is optional to get tagSNPs related information.

Fig. 8. The query results for interested drug target “Coagulation factor X” with “Preferred Name” and “Annotation”.

2.1.2.2. Metabolizing enzymes (Clicking arrow line 2 in Fig. 6)

The results show “Prothrombin” and “Vitamin K-dependent protein C” metabolizing enzymes information. Users can click the “Vitamin K-dependent protein C” (arrow line 2 in Fig. 6) to get further STRING [2] protein interaction information. When users click the “Vitamin K-dependent protein C” button, the waiting page for “STRING protein Query” will be shown as Fig. 7.

After waiting for a moment, the query results for interested metabolizing enzymes “Vitamin K-dependent protein C” which have “Preferred Name” and “Annotation” will be shown as Fig. 9. In this case, two interested genes (“PROC” indicated by arrow line 1 in Fig. 9 and “PROS1”) are the output. On the top of this page, the selection for tagSNPs of HapMap parameters, such as HapMap database version [3], population, pairwise methods, R square cut off, minor allele frequency (MAF) cut off, is optional to get tagSNPs related information.

Fig. 9. The query results for interested metabolizing enzymes “Vitamin K-dependent protein C” with “Preferred Name” and “Annotation”.

2.1.2.3. Drug transporters

In this example, no any available drug transporters are found. If there are available drug transporters for other examples, the operations are similar to 2.1.2.1. Drug targets and 2.1.2.2. Metabolizing enzymes.

2.1.2.4. Drug carriers

In this example, no any available drug transporters are found. If there are available drug carriers for other examples, the operations are similar to 2.1.2.1. Drug targets and 2.1.2.2. Metabolizing enzymes.

2.1.3. The query results for the interactive genes

To select the gene ‘PROC’ by clicking the button (arrow line 1 in Fig. 9), the waiting page for “STRING protein Query” will be shown in Fig. 7. After waiting for a moment, the query results for the interactive genes, such as PROC, SERPINA5, PROCR, F2R, GGCX, F5, PF4, SERPINC1, F8, THBD, and F3, to the selected gene (e.g., PROC), tagSNPs candidates (such as rs5936, rs1799810, rs2069916, and rs5937 for PROC; Please see the note mentioned in the legend of Fig. 10), PharmGKB info [4], and RFLP restriction enzymes (from REBASE) [5] availability with hyperlink are shown in Fig. 10. All these information is provided with hyperlinks.

Fig. 10. The query results for interactive genes, tagSNPs, PharmGKB, and RFLP restriction enzyme availability with hyperlinks. Caution: Tag SNP selection candidates from different operation times in HapMap may show different representative tagSNPs due to the greedy algorithm built, i.e., they will randomly select one tagSNP from one LD block. Some tag SNPs may or may not find in next test. Likely, Drug-SNPing has the same performance.

2.1.4. The query results for STRING interaction

In order to get STRING interaction, users can click the big image “STRING interaction” (indicated by an arrow line 1 in Fig. 10). The visualization for STRING interaction will be shown in Fig. 11. More or less interactions (indicated by arrow lines 4 and 5 in Fig. 11) are also available for adjustment as shown in Fig. 12.

Fig. 11. The visualization for the STRING interactions. The meanings of these arrow lines are explained later. Except the “color balls” (protein or gene name) are linked to our designed SNP genotyping path, the relationships (combined association score) between these color balls are online retrieved from the STRING [2].

Fig. 12. More or less interactions for STING-based SNP genotyping related information.

On the top of Fig. 11, two selection parameters, such as 1) NCBI SNPs (indicated by arrow line 1) and 2) HapMap tagSNPs (indicated by arrow line 2), are optional for the visual image to get SNP related information. They are described in detail as 4.2. NCBI SNP selection and HapMap tagSNPs. (Page 26)

2.2. In the example of FMO3 for HUGO gene name input

Using the “FMO3”, we demonstrate the improvement for gene-centric query for drugs which are mentioned as some drug targets. Clicking the arrow lines 1 and 2 in Fig. 13, the results are shown in Fig. 14.

Fig. 13. The gene name input for the ‘FMO3’ gene.

2.2.1. The query results for drug information

Clicking the arrow line 1 in Fig. 13, the original information for DrugBank is available such as DrugBank ID (DB00250). The hyperlink result is shown in Fig. 14. This drug information for DB00250 demonstrates that “FMO3” is one of the target genes.

Fig. 14. The query results for drug information which the input of gene name “FMO3”. Here, only part (ID1 to ID4) of the results is shown (total IDs are ID1 to ID8). In contrast, it fails to list this information by gene name (HUGO) input as mentioned in Fig. 15 and Fig. 16.

2.2.2. Gene name (HUGO) searching problem in DrugBank

Fig. 15. DrugBank information for DB00250 demonstrates that “FMO3” is one of the enzymes. DB00250 is the ID1 in Fig. 14.

However, the query strategy for the DrugBank seems to focus the gene search by drug name (such as the “Dapsone”) rather than the official gene name (HUGO) for searching. For example, the HUGO gene name- “FMO3” input in DrugBank (clicking the arrow line 1 “Search” button in Fig. 16) cannot find the drugs related to FMO3 gene (arrow line 2 in Fig. 16) which is demonstrated in Fig. 15.

Fig. 16. Query problem for gene name (HUGO; official gene symbol) input in DrugBank. In the example of FMO3 (HUGO gene name input), the DrugBank fails to find the drug information for DB00250 (http://www.drugbank.ca/search/search?query=FMO3) although the gene name is listed in its DrugBank information as shown in Fig. 15.

Similarly, DrugBank information for DB00132 and DB00155 demonstrates that FABP7 and PADI4 are some of the target genes as shown in Fig. 17 and Fig. 18, respectively. However, the DrugBank failed to find the drugs related to the FABP7 and PADI4 (Fig. 32) genes. Therefore, the DrugBank failed to search the drugs by HUGO gene name input in current version (access at 2012/10/20).

Fig. 17. DrugBank information for DB00132 demonstrates that “FABP7” is one of the target genes.

Fig. 18. DrugBank information for DB00155 demonstrates that “PADI4” is one of the target genes.

Fig. 19. Query problem for gene name (HUGO; official gene symbol) input in DrugBank. In the examples of FABP7 and PADI4 (gene name inputs), the DrugBank failed to find the target genes (http://www.drugbank.ca/search/search?query=FABP7 and http://www.drugbank.ca/search/search?query=PADI4) for drug information for DB00132 and DB00155 although the gene names are listed in its DrugBank information as shown in Fig. 17 and Fig. 18, respectively.

2.2.3. Gene name (HUGO) searching is improved in Drug-SNPing

In contrast, the Drug-SNPing is the gene-centric search for drugs with the gene name (HUGO) input. For example, clicking the arrow line 2 in Fig. 14, the target gene “FMO3” is successfully identified with pink color background (arrow line 1 in Fig. 20). The other non-FMO3 genes for the selected Drug ID DB00250 are also provided for reference (arrow line 2, 3, and 4 in Fig. 20) if users are interested in these genes.

Fig. 20. Drug-SNPing can solve the query problem for gene name (HUGO; official gene symbol) input in DrugBank in the example of FMO3 (gene name input). Other gene name inputs are also available in our proposed Drug-SNPing system (not shown in this figure). The Drug Targets (1) and (2) are the targets listed in DB00250. (Arrow line 1) The gene target “FMO3 (Dimethylaniline monooxygenase [N-oxide-forming] 3)” is highlighted in pink color for background as shown in Metabolizing Enzymes (3). (Arrow lines 2, 3, and 4) The other non-FMO3 gene of Drug Targets and Metabolizing Enzymes listed in DB00250. The Drug Transporters and Drug Carriers are not available in DB00250.

Due to the “FMO3” drug targets information is the preferred HUGO gene, users can click the “Dimethylaniline monooxygenase [N-oxide-forming] 3” (arrow line 1 in Fig. 20) to get further STRING protein interaction information. When users click the “Dimethylaniline monooxygenase [N-oxide-forming] 3” button, the waiting page for “STRING protein Query” will be shown as Fig. 21.

Fig. 21. The waiting page for “STRING protein Query” (The same as Fig. 7).

After waiting for a moment, the query results for interested genes containing “Preferred Name” and “Annotation” will be shown as Fig. 22. In this case, only one interested gene (FMO3) as indicated by arrow line 1 in Fig. 22 is output. To select the gene ‘FMO3’ by clicking the button (arrow line 1 in Fig. 22), the query results for interactive genes is shown in Fig. 23 (see 2.1.3. The query results for the interactive genes for details).

Fig. 22. The query results for interested gene with “Preferred Name” and “Annotation”. (Similar to Fig. 9)

Fig. 23. The query results for interactive genes, tagSNPs, and RFLP restriction enzyme availability with hyperlinks in the example of FMO3 gene. Caution: Tag SNP selection candidates from different operation times in HapMap may show different representative tagSNPs due to the greedy algorithm built, i.e., they will randomly select one tagSNP from one LD block. Some tag SNPs may or may not find in next test. Likely, Drug-SNPing has the same performance. (Similar to Fig. 10)

Return