NCI Database Interface FAQ

How can I link directly into your system? For example, I'd like to call up individual compounds from my own Web service.
Construct a URL such as `http:///ncidb2.2/nci2.2.tcl?op1=cas&data1=4187-87-5&output=table&nomsg=1` to call up a single compound in Detail Pane view by CAS RN ("op1=cas"); or `http:///ncidb2.2/nci2.2.tcl?op1=nsc&data1=4326&output=detail &highbondlist=1+2+3+4+5+6+7+8+9+10&highatomlist=1+2+3+4+5+6+7+8+9+10&conflist=-1&passid=&nomsg=1` to call up a single compound in Detail Pane view by NSC # ("op1=nsc"). (Note that the URL should be entered as one string in the web browser's location field.) Generally, you'll want to use "op1" as the search criteria, "method1" as a suboption (if applicable), and "data1" for the value of the input field of the form. The "nomsg=1" parameter must be set. Thus, Search by NSC Number: op1=nsc&data1=3215 Search by CAS RN: op1=cas&data1=1-2-3 Substructure search (SMILES string as input string): op1=ss&data1=CCC Name search: op1=name&method1=substring&data1=bla (other methods for name searches: exact, exactn, substringn, regexp, shell). This will call up the information as you would see it in the appropriate pane of our service, but will not preserve the frame structure. If this is important to you (e.g. because you want to allow the user to conduct further searches etc. within the Enhanced NCI Database Browser), you can call our service directly with an NSC number list with a URL contructed as in `http:///ncidb2.2/?nsc=10+50+99`
Can I share search results with others? I'd like to use your Web site in a collaborative project.
Absolutely. Conduct your search as desired. Export your hit list in the format Cactvs Hitlist (records). Send the saved file to the person you want to share it with. They can then use the Query Type Merge/Upload Hitlist in the bottommost Query row to re-import this hit list. If no other search fields are used, this list will simply be loaded; otherwise, the search will be conducted on this subset of structures.
Why are all NCI compounds identified by an "NSC" number and not an "NCI" number? What does "NSC" stand for?
The term NSC [number] is derived from (part of) the acronym of the Cancer Chemotherapy National Service Center (CCNSC). This organization, a predecessor of the DTP, was established at the NCI by virtue of a Congressional authorization in 1955 of $5 million for the establishment of an anticancer Drug Development Program, and became fully operational in September of 1957.
How do I specify that an atom in a substructure query should (not) be in a ring?
There are several methods to achieve this. The simplest way is to use the built-in Java editor. Just draw your structure, then open the query specification window by clicking on the QRY button, select the atom attribute (ring/nonring) and click on the atom in the drawing area. Note that by default the editor applet resets the atom type to any. You may still want to specify the element by also selecting an appropriate element symbol from the editor query panel. Alternatively, you can also use an external editor such as ISIS/Draw. Draw the query structure and store it in MDL Molfile format on your disk. Then use the upload button (on the fourth row on the query panel) to load your query specification into the database engine. We support, with very few exceptions, the full range of ISIS/Base query features, including 3D query attributes. The only feature which is currently unsupported is complex R-group logic - but simple R-group queries will work.
How do I restrict my search to structures which contain a certain substructure and show some anti-viral anti-AIDS activity?
First, draw your structure as usual and transfer it into one of the query data fields. Select Substructure Search for that field. Then take one of the yet unused data input rows and switch its query type to AIDS Screen Result. The menu below the data input field will change. Select from that menu the entry labelled Active. Make sure that the Connect query fields radiobutton is set to AND, so that only structures which match the desired AIDS screening result and the substructure are displayed. It does not matter which field you are using for the structure query, and which field for the AIDS screening result - the software will always re-order the query to achieve maximum performance (meaning that the AIDS screening results is always checked first).
I got so many results that the maximum number of hits was exceeded. I still want to see those undisclosed matches.
This is easy. On the hitlist panel, the third row of control elements, which is labelled Miscellaneous, contains an entry Retrieve Next Hits whenever the end of the database file was not reached during the query. If you select this option, and press the button to the right of the menu, the current query will be re-executed starting at the file position where you stopped. If the menu entry is not present, the query reached the end of the file and no more hits will be found. In any case, you can restart the query from the beginning of the file by choosing Restart Query from the menu.
I selected Sort by Similarity, and nothing happens!
Sorting by similarity has an effect only if your query contains at least one similarity query part, i.e. one query field must be Similarity Search. You need to have some full structure to compare the hits against, so sorting by similarity when your query is Give me NSC numbers 1-10 is meaningless. Computing the similarity to a substructure is an operation of rather dubious value. In the future, we may add some clustering facility to group matches.
I cannot understand why Nickel(II) 4-cyclohexylbutyrate (NSC number 7729) has the formula C₁₀H₁₈O₂ and does not contain nickel?
We have verified that this data item is already missing the nickel atom in the original structure files used to build the database, so this is not a software problem on our side. After more detailed analysis, it seems that the nickel atom is actually in the central NCI registry, but it is lost when the structure was exported. There are other known problems with salts and metal-counterion compounds which are currently not dumped correctly by NCI's software in its export format. The Nickel atom is also missing on the DTP raw data file. We have been promised a fix for two years now, but nothing has happened.
I want the full database as SD-file/specially formatted SMILES dump etc. for my in-house system!
There are essentially four ways of achieving this. First, you can download selected subsets of the database in various formats yourself, with user-selectable additional data fields, in blocks of up to 10,000 compounds when done interactively, by exporting structure lists as SD-file etc. from the hitlist pane. Second, you can download the raw files used to build this service, including the raw SMILES structure file (which has a still manageable file size), in bulk form from our Download page. Third, if this is not sufficient, you can ask us for a custom conversion. Note however that, while the basic data is free and can, in many small and unlinked pieces, be downloaded from the NCI site (http://dtp.nci.nih.gov/webdata.html) and some other places, we feel it is not our duty to perform free conversion services for undisclosed industry projects. If you are working on an university research project, give us credits and make it plausible that the Web-based system does not fulfill your needs, we are certainly willing to help. If you are from industry, please be prepared to pay a reasonable fee for our time and efforts [this applies only to W. D. Ihlenfeldt]. But, please, again note that the underlying data used to build this service will always be freely available from NCI, and that we would only request compensation for our work. Given the first three possibilities, the fourth way should never be necessary: We do not appreciate it if you write robot scripts to automatize the download by interaction with the Web server, especially if every robot contact only retrieves a single structure and involves a complex, time-consuming query. Hosts which are the source of such activity will be locked out.

How can I link directly into your system? For example, I'd like to call up individual compounds from my own Web service.

Construct a URL such as


http:///ncidb2.2/nci2.2.tcl?op1=cas&data1=4187-87-5&output=table&nomsg=1

to call up a single compound in Detail Pane view by CAS RN ("op1=cas"); or


http:///ncidb2.2/nci2.2.tcl?op1=nsc&data1=4326&output=detail
&highbondlist=1+2+3+4+5+6+7+8+9+10&highatomlist=1+2+3+4+5+6+7+8+9+10&conflist=-1&passid=&nomsg=1

to call up a single compound in Detail Pane view by NSC # ("op1=nsc"). (Note that the URL should be entered as one string in the web browser's location field.)
Generally, you'll want to use "op1" as the search criteria, "method1" as a suboption (if applicable), and "data1" for the value of the input field of the form. The "nomsg=1" parameter must be set. Thus,

Search by NSC Number: op1=nsc&data1=3215
Search by CAS RN: op1=cas&data1=1-2-3
Substructure search (SMILES string as input string): op1=ss&data1=CCC
Name search: op1=name&method1=substring&data1=bla
(other methods for name searches: exact, exactn, substringn, regexp, shell).

This will call up the information as you would see it in the appropriate pane of our service, but will not preserve the frame structure. If this is important to you (e.g. because you want to allow the user to conduct further searches etc. within the Enhanced NCI Database Browser), you can call our service directly with an NSC number list with a URL contructed as in


http:///ncidb2.2/?nsc=10+50+99

Can I share search results with others? I'd like to use your Web site in a collaborative project.

Absolutely. Conduct your search as desired. Export your hit list in the format Cactvs Hitlist (records). Send the saved file to the person you want to share it with. They can then use the Query Type Merge/Upload Hitlist in the bottommost Query row to re-import this hit list. If no other search fields are used, this list will simply be loaded; otherwise, the search will be conducted on this subset of structures.

Why are all NCI compounds identified by an "NSC" number and not an "NCI" number? What does "NSC" stand for?

The term NSC [number] is derived from (part of) the acronym of the Cancer Chemotherapy National Service Center (CCNSC). This organization, a predecessor of the DTP, was established at the NCI by virtue of a Congressional authorization in 1955 of $5 million for the establishment of an anticancer Drug Development Program, and became fully operational in September of 1957.

How do I specify that an atom in a substructure query should (not) be in a ring?

There are several methods to achieve this. The simplest way is to use the built-in Java editor. Just draw your structure, then open the query specification window by clicking on the QRY button, select the atom attribute (ring/nonring) and click on the atom in the drawing area. Note that by default the editor applet resets the atom type to any. You may still want to specify the element by also selecting an appropriate element symbol from the editor query panel.

Alternatively, you can also use an external editor such as ISIS/Draw. Draw the query structure and store it in MDL Molfile format on your disk. Then use the upload button (on the fourth row on the query panel) to load your query specification into the database engine. We support, with very few exceptions, the full range of ISIS/Base query features, including 3D query attributes. The only feature which is currently unsupported is complex R-group logic - but simple R-group queries will work.

How do I restrict my search to structures which contain a certain substructure and show some anti-viral anti-AIDS activity?

First, draw your structure as usual and transfer it into one of the query data fields. Select Substructure Search for that field. Then take one of the yet unused data input rows and switch its query type to AIDS Screen Result. The menu below the data input field will change. Select from that menu the entry labelled Active.

Make sure that the Connect query fields radiobutton is set to AND, so that only structures which match the desired AIDS screening result and the substructure are displayed. It does not matter which field you are using for the structure query, and which field for the AIDS screening result - the software will always re-order the query to achieve maximum performance (meaning that the AIDS screening results is always checked first).

I got so many results that the maximum number of hits was exceeded. I still want to see those undisclosed matches.

This is easy. On the hitlist panel, the third row of control elements, which is labelled Miscellaneous, contains an entry Retrieve Next Hits whenever the end of the database file was not reached during the query. If you select this option, and press the button to the right of the menu, the current query will be re-executed starting at the file position where you stopped. If the menu entry is not present, the query reached the end of the file and no more hits will be found. In any case, you can restart the query from the beginning of the file by choosing Restart Query from the menu.

I selected Sort by Similarity, and nothing happens!

Sorting by similarity has an effect only if your query contains at least one similarity query part, i.e. one query field must be Similarity Search. You need to have some full structure to compare the hits against, so sorting by similarity when your query is Give me NSC numbers 1-10 is meaningless. Computing the similarity to a substructure is an operation of rather dubious value. In the future, we may add some clustering facility to group matches.

I cannot understand why Nickel(II) 4-cyclohexylbutyrate (NSC number 7729) has the formula C₁₀H₁₈O₂ and does not contain nickel?

We have verified that this data item is already missing the nickel atom in the original structure files used to build the database, so this is not a software problem on our side. After more detailed analysis, it seems that the nickel atom is actually in the central NCI registry, but it is lost when the structure was exported. There are other known problems with salts and metal-counterion compounds which are currently not dumped correctly by NCI's software in its export format. The Nickel atom is also missing on the DTP raw data file. We have been promised a fix for two years now, but nothing has happened.

I want the full database as SD-file/specially formatted SMILES dump etc. for my in-house system!

There are essentially four ways of achieving this.

First, you can download selected subsets of the database in various formats yourself, with user-selectable additional data fields, in blocks of up to 10,000 compounds when done interactively, by exporting structure lists as SD-file etc. from the hitlist pane.

Second, you can download the raw files used to build this service, including the raw SMILES structure file (which has a still manageable file size), in bulk form from our Download page.

Third, if this is not sufficient, you can ask us for a custom conversion. Note however that, while the basic data is free and can, in many small and unlinked pieces, be downloaded from the NCI site (http://dtp.nci.nih.gov/webdata.html) and some other places, we feel it is not our duty to perform free conversion services for undisclosed industry projects. If you are working on an university research project, give us credits and make it plausible that the Web-based system does not fulfill your needs, we are certainly willing to help. If you are from industry, please be prepared to pay a reasonable fee for our time and efforts [this applies only to W. D. Ihlenfeldt]. But, please, again note that the underlying data used to build this service will always be freely available from NCI, and that we would only request compensation for our work.

Given the first three possibilities, the fourth way should never be necessary: We do not appreciate it if you write robot scripts to automatize the download by interaction with the Web server, especially if every robot contact only retrieves a single structure and involves a complex, time-consuming query. Hosts which are the source of such activity will be locked out.

The NCI Database Interface FAQ