To perform a chemical structure search you must provide an appropriate query.  There are several forms of structure input allowed:

 

• Manually Drawn Structure Input
  A chemical structure query may be manually entered or edited using the PubChem Sketcher.

 

• SMILES/SMARTS Input
  SMILES -- Simplified Molecular Input Line Entry System, a chemical structure line notation (a typographical method using printable characters) for entering and representing molecules. SMILES strings can be imported or exported from many molecular editors. SMILES can include wildcard atoms using a "*" for the atom. For example, the SMILES of gleevec is CC1=C(C=C(C=C1)NC(=O)C2=CC=C(C=C2)CN3CCN(CC3)C)NC4=NC=CC(=N4)C5=CN=CC=C5.

  SMARTS -- SMILES ARbitrary Target Specification, a chemical structure query line notation.  SMILES is a subset of SMARTS. As such, a valid SMILES is a valid SMARTS. All standard SMILES features, including stereochemistry and isotope labeling, and all standard SMARTS extensions, including recursive SMARTS, are recognized.

   

• InChI Input
  InChI -- IUPAC International Chemical Identifier, a chemical structure line notation. For example, the InChI string of aspirin is InChI=1/C9H8O4/c1-6(10)13-8-5-3-2-4-7(8)9(11)12/h2-5H,1H3,(H,11,12)/f/h11H.

   

• CID Input
  CID -- PubChem Compound Identification, a non-zero integer PubChem accession identifier for a unique chemical structure. For example, the CID of aspirin is 2244.

   

• Molecular Formula Input
  Molecular Formula (MF) -- Specification of the count of each element in a compound. For example, ethyl alcohol has the MF C₂H₆O.

  The general MF query syntax consists of a series of valid atomic symbols (please consult your periodical chart), each optionally followed by either a number or a range. An "*" may be used as a wildcard in the place of a number. The generic range syntax is "[atomic symbol][low count]-[high count]", repeated for every specified element. Elements may be written in arbitrary order.

  Examples:

  1. C7-8:	represents compounds with seven or eight carbons.
  2. C-7:	represents compounds with up to seven carbons.
  3. C7-:	represents compounds with seven or more carbons
  4. C or C1:	represents compounds with exactly one carbon
  5. C-:	represents any number of carbons, including none.

  Please note that two-letter elements must be written with the second letter in lowercase, otherwise "Cu" (one copper) and "CU" (one carbon, one uranium) would not be distinguishable.

   

• File Input
  A chemical structure file or a molecular formula file on your computer may be uploaded as a query. For multiple molecular formulas, this is one MF query per line.  For multiple chemical structures in a file, all structures must be of the same format (e.g., mixing of SMILES/SMARTS and InChI is not allowed).  If a line notation is being used, only one chemical structure or chemical structure query is allowed per line.  "SDF" formatted files (both v2000 and v3000) are also allowed. While chemical structure formats other than those listed above may work, they are not supported. When multiple queries within the same file are provided, the final result is an "OR" of each query result.

PubChem Structure Search provides various chemical search types and options. On the top of structure search page there are several tabs. Each of these tabs is dedicated to a particular search type.  Click one of the tabs to perform the desired search type.

Search By:

Name/ Text

Identity/ Similarity

Substructure/ Superstructure

Molecular Formula

3D Conformer

Saved Search

To search the PubChem Compound Database using a textual query (and not by a chemical structure query), please click the Name/Text tab to go to the appropriate query form:

A Name/Text query allows one to locate chemical structures using one or more textual keywords. For example, synonyms, descriptors, or MeSH terms can be entered for searching. To search for a phrase or a name with non-alphanumeric characters (i.e., not of a-z, A-Z, or 0-9), please use double quotes around the query (for example, "4-guanidinobutyramide"). There are various indexes that can be individually searched by suffixing the text phrase or keyword with an appropriate index and by using brackets (for example, to search a particular IUPAC name: "4-(diaminomethylideneamino)butanamide"[iupacname]). One may also perform numeric range searches of appropriate index fields using a ":" delimiter (for example: 100.5:200[molecularweight] for a molecular weight range search between 100.5-200.0 g/mol). Queries may be combined using the Boolean operators "AND", "OR", and "NOT".  Please note that these Boolean operators must be capitalized.

Examples:

• Which compound has a CID equal to 123?
  
     Enter 123[cid] or 123[uid] in the text box, and click the Search button.
  
  

  	Synonyms/Descriptors/MeSH term etc:     Search

  
  
  
• Which compounds have a molecular weight between 100 and 200 daltons?
     Enter they query 100:200[molecularweight] or 100:200[mw] in the text box, and click the Search button.
  
  

  	Synonyms/Descriptors/MeSH term etc:     Search

  
  
  
• Which compounds have "aspirin" or "gleevec" as keywords in their synonyms?
      Enter aspirin[synonym] OR gleevec[synonym] in the text box, and click Search button.
  
  

  	Synonyms/Descriptors/MeSH term etc:  Search

To locate a particular chemical structure or to find chemical structures similar to another chemical structure, please click the Identity/Similarity tab to go to the appropriate query form:

All queries for identity search and similarity search must be valid chemical structures and will be subjected to the same PubChem chemical structure standardization processing used to create the PubChem Compound database from the PubChem Substance database.  This "standardization" procedure will normalize the chemical structure query for resonance and tautomeric form invariance to make it consistent with its PubChem representation, thus increasing the likelihood of finding a particular chemical structure in PubChem.  Invalid or incompletely defined chemical structure queries may not provide intended results.

Please note that if explicit hydrogen atoms are not provided in the chemical structure query, PubChem will use the most likely valence state for the atoms in the provided chemical structure. To achieve desired and consistent results, it is best to provide all explicit hydrogens in chemical structure queries of this search type.

• Identical Structures search allows you to locate records that are identical to the provided chemical structure with different notions of chemical structure identity. Additional search options allow you to choose the degree of "sameness":
  • Any Tautomer provides tautomeric form invariance and returns chemical structures that are tautomers of the provided chemical structure. This option ignores stereochemical or isotopic differences.

  • Same Connectivity provides isotopic and stereocenter form invariance and returns chemical structures that exactly match the connectivity and valence-bond form (i.e., same bond orders) of the provided chemical structure. This option ignores differences in isotopic or stereochemical information.

  • Same Stereoisomer provides isotopic form invariance and returns chemical structures that exactly match the connectivity, valence-bond, and stereochemical form of the provided chemical structure. This option ignores isotopic differences.

  • Same Isotopic Labels provides stereochemical form invariance and returns chemical structures that exactly match the connectivity, valence bond, and isotopic form of the provided chemical structure. This option ignores stereochemical differences.

  • Same Stereochemistry and Isotopes provides chemical structures that exactly match the connectivity, valence bond, stereochemical, and isotopic form of the provided chemical structure.  This option should be used in most cases and can be considered an "exact match" to a chemical structure.

  • Same Isotopes and Nonconflicting Stereoisomers returns chemical structures that exactly match the connectivity, valence bond, and isotopic form but allows either the same or unspecified stereo at each defined stereocenter in the provided chemical structure query.

  • Nonconflicting Stereoisomers provides isotopic form invariance and returns chemical structures that exactly match the connectivity and valence bond form but allows either the same or unspecified stereo at each defined stereocenter of the provided chemical structure query.
  
  
• Similar Compounds search type allows you to locate records that are similar to a chemical structure query using pre-specified similarity thresholds. Similarity is measured using the Tanimoto equation and the PubChem dictionary-based binary fingerprint. This fingerprint consists of series of chemical substructure "keys". Each key denotes the presence or absence of a particular substructure in a molecule. The fingerprint does not consider variation in stereochemical or isotopic information. Collectively, these binary keys provide a "fingerprint" of a particular chemical structure valence-bond form.

  The degree of similarity is dictated by the Threshold parameter.  A threshold of "100%" effectively acts as an "exact match" to the provided chemical structure query (ignoring stereo or isotopic information), while a threshold of "0%" would return all chemical structures in the PubChem Compound database. Various predefined thresholds between 99-60% are allowed.

To work properly, both Identity and Similarity search types require a completely defined structure. A SMILES string with wildcards or a SMARTS string containing substructure patterns may not provide intended results for Identity and Similarity search types.

To perform an Identity or Similarity search, please click the Identity/Similarity tab in the Search By section to choose the above described search types and method for providing a chemical structure:

• To draw a chemical structure, click the Launch button to open the PubChem Sketcher.
  
  
  
• To provide a chemical structure by CID, SMILES or InChI click on the CID, SMILES, InChI tab.
  
  
  
• To Upload a structure file (e.g., in SDF format), click the Structure File tab.
  
  
  

Finally, select a desired option for your search job.

• The default search option is "Identical Structures" with several choices for finding "Similar Compounds".
  
  
  
  
  
• By clicking on to expand the Option section, more options will appear for Identity search.
  
  
  
  
• By clicking on to expand the Option section, more options will appear for Similar Structure search.
  
  
  
  

A number of filters are also available under "Filter" section.

When an Identity/Similarity search is submitted, the provided structure will be validated and standardized, prior to performing the query.

To locate a chemical structures using a particular chemical structure pattern, please click the Substructure/Superstructure tab to go to the appropriate query form:

The presence of explicit hydrogen atoms can be a complicating factor for substructure and superstructure queries, considering the hydrogen atoms become part of the pattern being searched.  When a CID is provided as the structure input method, by default explicit hydrogens are removed prior to search.

• Substructure search allows one to locate chemical structures that contain a particular connectivity and valence-bond (i.e., bond order) pattern. For example, a substructure search of ethanol (SMILES: OCC) would return, among others, acetic acid (SMILES: OC(=O)C), since ethanol is a substructure of acetic acid.
  
  
• Superstructure search allows one to identify chemical structures that comprise or make up (i.e., is a substructure of) the provided chemical structure query. For example, a superstructure search of acetic acid (SMILES: OC(=O)C) would return, among others, ethanol (SMILES: OCC), since ethanol is a substructure of acetic acid.

  There are additional matching options for both the Substructure and Superstructure searches. These options are provided for further flexibility within a structural query.

  • Match stereochemistry
    The pull-down menu for matching stereochemistry allows users to specify the type of stereo matching:
    Ignore stereochemistry in the query;
    Exact stereochemistry (i.e., allows only the enantiomer provided in the query);
    Relative stereochemistry (i.e., allows all enantiomers);
    Nonconflicting stereochemistry (i.e., allows the enantiomer or unspecified stereo at each defined stereocenter in the query).

  • Match isotope
    Indicates that the isotopic information within the query is to be used. By default, isotopic information is ignored.

  • Match charge
    Indicates that the formal charge information within the query is to be used. By default, formal charge information is ignored.

  • Match tautomers
    Indicates that tautomers of the query are allowed to match.  Please note that the structure must be capable of having a tautomer for this option to provide any benefit.

  • Ringsystems not embedded
    Indicates that the results should exactly match the ring systems as they are provided in the query.  Specifying this option means that no part of the query can be embedded in a larger ring system.

  • Single/double bonds match aromatic bonds
    Indicates that single or double bonds in the query can match aromatic bonds.

  • Chain bonds match ring bonds
    Indicated that chain bonds in the query can match both ring and chain bonds.

  • Strip hydrogen
    Indicates that explicit hydrogen in the query is to be removed prior to searching. Explicit hydrogens are not automatically added for substructure searches. However, adding explicit hydrogens to all sites where you do not want any substituents will both focus your search and speed it up.

To perform a Substructure or Superstructure search, please click the Substructure/Superstructure tab in the Search By section and choose one the above described search input type tabs. Same as for Identity and Similarity search types, one can draw a chemical structure or provide a chemical structure by CID, SMILES/SMARTS, InChI, or upload a structure file (e.g., in SDF format).

Finally, select a desired search type option for the search job.

• Default search options allow: single or double bonds in the query can match aromatic bonds, chain bonds match both ring and chain bonds, and, when CID is the input type, all explicit hydrogen atoms in the query are automatically stripped.
  
  
  
  
  
• By clicking to expand the Option section more options will appear for these search types.
  
  

When an Substructure/Superstructure search is submitted, the provided structure will be displayed as an image for user's preview.

To locate a chemical structures by a particular molecular formula pattern, please click the Molecular Formula tab to go to the appropriate query form:

 

Molecular Formula search allows one to locate records that contain a certain number and type of atomic elements. For example, a query of "C₆H₆" will return all structures containing six (6) carbon atoms and six (6) hydrogen atoms.

Molecular Formula search is specifically designed for queries that are either a single molecular formula or a molecular formula file, containing one or more molecular formulas, one per line. A CID may also be used to provide the input molecular formula.

By default, the search results from a Molecular Formula search will exactly match the entered stoichiometry. One may allow other elements in the returned results by click on the following the Options, and select "allow other element".
 

To locate a chemicals structurally similar to a 3D conformer, please click the 3D Conformer tab to go to the appropriate query form:

The 3D conformer search can take either a 2D or 3D chemical structure as input. A 3D structure may be submitted only through a SDF. The search database consists of pre-calculated 3D conformers of compounds in the PubChem Compound database. Currently, up to 3 conformers per compound are in the search database. A 3D conformer search job will look for the chemicals in the PubChem Compound database that have 3d conformers structurally similar to the query conformer. The threshold of the shape similarity is 80%, and of the feature similarity is 50%.

When the input is a 2D structure, if the input is a CID, the pre-calculated default conformer of that CID is chosen to search for similar conformers; otherwise, the theoretical 3D conformers of the query will be generated by OpenEye software Omega, and only a single conformer withenergy-minimum is utilized as the query conformer. A 3D input structure will be directly used to search for similar conformers.

The search results can be sorted. The pull-down menu in the Option section allows you to specify the sorting of search results:

The search results can be sent into the PubChem web-based 3D conformer superposition viewer to understand the details of the similarities and superpositions between similar conformers. They can also be sent and displayed into NCBI Entrez search engine for further investigations via utilization of various NCBI scientific resources. Choosing a destination from the "Output to: " pull-down menu in the Option section to decide where your search results will be displayed:

PubChem Structure Search allows queries to be imported and exported via PUG XML.  These saved queries can be sent to colleagues or used to repeat a search at a later date without having to reenter the query or options.

After entering a chemical structure and selecting appropriate options for Identity/Similarity, Substructure/Superstructure, or Molecular Formula search, you have the option of saving your query. This is achieved by clicking Save Query button at the bottom of the corresponding submission page. The saved query file is in XML format based on the PUG XML schema for PubChem Structure Search. This saved query can be also used as input for the PUG service.

Users can directly upload a saved query file for repeating a previously saved search. To upload the saved search, one needs to click on the Saved Query tab in the Search By section, and click on Browse button on the page like the one below to select a file for upload.

It is possible to modify a saved query by clicking the Edit Query button depicted in the image above. Data from the saved search file will be used to fill the query form. In that form, one can modify any settings so that the query is modified. For example, if a query file contains the data for an 80% similarity search of C1CCCCC1, after you input the query file and click the Edit Query, a query form like the one below will be ready for you to change your query.

By default, there is no time limit for a structure search to be performed. A time limit can be imposed on the duration of chemical structure query. One may modify this limit by clicking the following the Option on a submission page. If the query does not finish within allowed time period, all hits found up to that time period (and within the hit limit specified), will be returned.

By default, the maximum number of results returned by the chemical structure query is currently limited to two million. One may modify this limit by clicking the following the Option on a submission page.

For example, these are what the Time Limit and Result Limit option interface looks like for an Identity search.

For Identity/Similarity and Substructure/Superstructure searches, submitted queries are first validated. Validation, in PubChem terminology, includes standardization processing of Identity/Similarity queries, much like the processing of chemical structures used to create the PubChem compound database from depositors' original structures. Structure Search Query Preview allows you to review the structure query being submitted to PubChem. This step provides an image of each query structure. Clicking the "Continue" button will allow the search job to proceed. To revise or resubmit a query, please click the "Review" button.

Here is a typical Structure Search Query Preview:



When multiple chemical structure queries are provided, these navigation buttons (first, previous, next, last, respectively):       enable you to review each. Below is a Query Preview of an SDF file for a similarity search containing one invalid chemical structure and two valid chemical structures. The number "#2/3" means the structure currently displayed is the 2^nd structure of three in the submission (#n^th/total). The Query Preview may contain one or more message containing structure query validation information.

Results of a structure search may be further narrowed using various filters. By clicking following the Filters section, one may select from many filter types. The list of filter types is presented below:

• Compound Subset
  One may search an explicitly specified subset of the PubChem Compound database. After selecting the desired subset type, one can specify any subset-specific option. There are three subset types available:
  
  • PubChem Compound Search History
    One may use a previously performed chemical structure query or PubChem Compound search result for further refinement. Click Retrieve to obtain a current search history list, if none is present. If a search history result is found (or if one already exists), a pull-down menu will appear offering to select a particular history item. Click Refresh to refresh available history items. Please keep in mind that a search history expires after eight hours of inactivity.
    
    
  • CID List
    A CID (PubChem Compound ID) list may be provided for subsetting a search. To search using a list of CIDs, select the CID List radio button, and input (for example, via copy/paste from a document) a list of CIDs.  One may use whitespace (e.g., a space, newline, etc.), a comma (","), or semicolon (";") as a delimiter.
    
    
  • CID File
    User can upload a text file containing a particular CID list. To search using such a CID file, select the CID File radio button, and browse to find the file that contains a list of CIDs (again, separated by either whitespace, comma (","), or semicolon (";") as a delimiter) on your computer.
    

• Property
  PubChem computes various properties of chemical structures.  The property filter allows you to restrict the search to chemical structures with particular computed property ranges. All ranges must be bounded. To bound an open-ended range, please enter a large value in the open-ended part of the text box, e.g., -999999 for a minimum value or 999999 for a maximum value.
  • MW -- Calculated molecular weight in units g/mol.
  • XLogP -- Log P calculated using XLogP method
  • HBD -- Count of hydrogen bond donors
  • HBA -- Count of hydrogen bond acceptors
  • Rotatable Bond -- Count of rotatable bonds
  • TPSA -- Polar surface area calculated using topological polar surface area method
  • Heavy Atom -- Count of heavy atoms, i.e., being those other than hydrogen atoms
  • Isotope -- Count of atoms with specified isotopic atom labels
  • Tautomer -- Count of unique tautomeric forms (to a maximum of 1,000).
  • Covalent Unit -- Count of covalently bonded moieties within a CID
  • Complexity -- Measure of structural complexity
  • Charge -- Total formal charge

• Stereochemistry
  Used for restricting search to chemical structures with particular stereochemical specificity, when stereochemistry is possible. Pull-down menus are available for specifying chiral centers and E/Z bonds:
  • Fully Specified -- All stereochemical information must be fully specified
  • Partially Specified -- All stereochemical information must be partially specified
  • Fully Unspecified -- All stereochemical information must be fully unspecified
  • Not Fully Specified -- All stereochemical information must not be fully specified
  • Not Partially Specified -- All stereochemical information must not be partially specified
  • Not Fully Unspecified -- All stereochemical information must not be fully unspecified

• BioActivity
  Used for restricting search to chemical structures with particular bioactivity. Pull-down menu is available to specify the type of information available in PubChem BioAssay with the following choices:
  • Tested -- Compound has a PubChem BioAssay result
  • Active -- Compound is active in a PubChem BioAssay
  • Inactive -- Compound is inactive in a PubChem BioAssay
  • Not Tested -- Compound does not have a PubChem BioAssay result

• Links
  Used for restricting search to chemical structures with particular cross-link types. User can filter the type of links for a compound by selecting the appropriate radio button with the choices by column:
  • Require -- Compound must have a link of this type
  • Disallow -- Compound must not have a link of this type
  • Allow -- Compound can have a link of this type

• Chemical Elements
  Used for restricting search to chemical structures containing particular chemical elements. Please check the boxes of the elements that must be present in the result set.

• Depositor's Category
  Used for restricting search to chemical structures matching a particular depositor category classification. User can filter the type of depositor category for a compound by selecting the appropriate radio button with the choices by column:
  • Require -- Compound must have a depositor category of this type
  • Disallow -- Compound must not have a depositor category of this type
  • Allow -- Compound can have a depositor category of this type

• Data Source
  Used for restricting search to chemical structures deposited by certain depositor(s). User can specify one or more data sources that are allowed or disallowed by selecting individual depositors in the scrollable text boxes:
  • From -- Compound resulted from data provided by selected depositors
  • Not from -- Compound did not result from data provided by selected depositors

PubChem Structure Search can be used directly without the need to fill out an input form.  This is achieved by formulating an appropriate URL with the input already provided.  The URL consists of a base (see below), being the same as that required to invoke the PubChem Structure Search with a "?cmd=search" at the end.  There are a series of parameters that may be provided to indicate the query type (q_type), the input query (q_data), and the search type.  The parameters are described in more detail below. 

Formulating the URL for a structure search is rather straightforward.  Each parameter may be provided in any order and must have "=" after the parameter name and a "&" character between parameters.  All query data input must be properly URL-encoded to enable proper interpretation.

• Base URL: https://pubchem.ncbi.nlm.nih.gov/search/
  
  
• URL parameters:
  • Query data type (q_type):  
    • single CID, InCh, or SMILES: q_type=dt
    • molecular formula: q_type=mf
    • structure file: q_type=str_file
    • molecular formula file: q_type=mf_file
    • saved query: q_type=xml
    
    
  • Query data (q_data):
    Combined with q_type, user can enter a CID, SMILE/SMARTS or InChI string, or a molecular formula by using this parameter. For example, to input CID 123, user has to give "q_type=dt&q_data=123" as part of your URL; to input a molecular formula "C₆H₆", user has to give "q_type=mf&q_data=C6H6" as part of your URL.
    
    
  • Search type (simp_schtp):
    • Identity search: simp_schtp=fs (means full structure)
    • Similarity search: simp_schtp=(%similarity) (for example, simp_schtp=90 means a 90% similarity structure search)
    • Substructure search: simp_schtp=subsch
    • Superstructure search: simp_schtp=supsch
    • Molecular Formula search: simp_schtp=mf

Examples:

• How to make an URL for a "Identity" search of CID 123?
  https://pubchem.ncbi.nlm.nih.gov/search/search.cgi?cmd=search&q_type=dt&q_data=123&simp_schtp=fs
  
  
• What is the URL for a 80% "Similarity" search of SMILES C1CCCCC1?
  https://pubchem.ncbi.nlm.nih.gov/search/search.cgi?cmd=search&q_type=dt&q_data=C1CCCCC1&simp_schtp=80
  
  
• How to program an URL for a "Substructure" search of an InChI string of aspirin: InChI=1/C9H8O4/c1-6(10)13-8-5-3-2-4-7(8)9(11)12/h2-5H,1H3,(H,11,12)/f/h11H?
  https://pubchem.ncbi.nlm.nih.gov/search/search.cgi?cmd=search&q_type=dt&q_data=InChI%3D1%2FC9H8O4%2Fc1-6(10)13-8-5-3-2-4-7(8)9(11)12%2Fh2-5H%2C1H3%2C(H%2C11%2C12)%2Ff%2Fh11H&simp_schtp=subsch
  
  
• What is a URL for a "Superstructure" search of carbonyl group representated by a SMARTS string [CX3]=[OX1]?
  https://pubchem.ncbi.nlm.nih.gov/search/search.cgi?cmd=search&q_type=dt&q_data=%5BCX3%5D%3D%5BOX1%5D&simp_schtp=supsch
  
  
• How to do a "Molecular Formula" search for "C₆H₆"?
  https://pubchem.ncbi.nlm.nih.gov/search/search.cgi?cmd=search&q_type=mf&q_data=C6H6&simp_schtp=mf
  
  Note: the quotation marks (") in the URLs above are unnecessary but for clarifying the data value of q_data, unless the input data contains a space or spaces.