Chemical Literature (Chem 184/284)

You are here: Home > Library Services > Library Instruction > Chemical Literature (Chem 184/284) > Lecture 18 Extra

Lecture 18 Extra: Why You Still Need to Go Online:
Specialized Techniques and Databases

While SciFinder Scholar and Scifinder Web are extremely powerful, the underlying command language, Messenger used in searching STN Online and STNWeb, is even more powerful. Of course like any command-driven interface, there is a much more difficult learning curve than in a menu-driven interface like SciFinder. So most users who are not information professionals prefer SciFinder.
Also, searching the STN interfaces uses a pay-by-the-search pricing metric, with fees depending on the number and types of searches done, how long you are connected to the database(s) and the number and format of results displayed. For end-user searching, academic institutions greatly prefer to have a flat-rate annual subscription. It can be pricy, but the costs are predictable, and the service can be widely promoted without worrying about someone accidentally running up a huge bill. Chemical Abstracts Service currently only offers flat-rate pricing for SciFinder for academic institutions, not for the command language versions. So, we've gone with SciFinder.
Nonetheless, there are certain searches that can only be done in the command language version, so the Library maintains an account with STN so I can do these searches as needed.

Searching for Special Substances in the Registry File

Certain classes of compounds pose special challenges for SciFinder searching -- they don't have well-defined stoichiometry, or may be defined in unusual ways, or are too large for standard structure searching.
CAS has implemented special techniques for some of these classes in the parent Registry File using the Messenger command language which are not yet available in SciFinder. Some of the specialized information for these substances is displayable, but not searchable, in SciFinder Scholar.

Chemical Names

Generally speaking, Chemical Name searching in SciFinder searches the exact chemical name which you enter. Name searching with the Messenger command language is more flexible.
The Basic Index of the Registry File consists of chemical name fragments from the Index Name (/in) and other chemical names (/cn), plus molecular formulas and Registry Numbers.
Proximity operators and right-hand truncation may be used in the Basic Index.
Chemical Name Fragmentation Example
dicyclopenta[def,pqr]tetraphenylene-1,8-dione
breaks into:
- dicyclopenta (plus di, cyclo, penta, dicyclo, cyclopenta)
- [def,pqr]
- tetraphenylene
- 1,8 (plus 1 and 8)
- dione (plus di and one)
Trade names and the like are usually not fragmented (except at spaces and punctuation).
Combinations of name fragments (using the (L) operator) can be very powerful for searching if you are not certain of the full name of a compound, but can make an educated guess about its nomenclature. Use with molecular formula (/MF) searching.

Chemical Name Field (/cn)

This index contains the complete chemical name from the IN (index name) field as well as all the synonyms in the record (e.g. trade names, common names and old CA index names.)
Note that not all possible synonyms are listed, and many compounds are listed only under their systematic names.
Remember that stereoisomers, mixtures, etc. each get their own Registry Numbers.
The EXPAND command in STN databases lets you browse specific indexes (similar to, say, the "Title begins with" search in PEGASUS.) EXPAND is very useful in /cn searches; it lets you check the form of a name without wasting money on incorrect search terms.
Remember that the E#'s appear in strict left to right alphanumeric order (and that numbers come before letters, and spaces and other punctuation before either.)

Examples:


=> e benzene, 1,2-dimethyl-/cn

E1           1     BENZENE, 1,2-DIMETHYL(OCTADECYLOXY)-/CN
E2           1     BENZENE, 1,2-DIMETHYL(PHENYLETHYL)-/CN
E3           1 --> BENZENE, 1,2-DIMETHYL-/CN
E4           1     BENZENE, 1,2-DIMETHYL-, ALUMINUM COMPLEX/CN
E5           1     BENZENE, 1,2-DIMETHYL-, ALUMINUM-TITANIUM COMPLEX/CN
E6          14     BENZENE, 1,2-DIMETHYL-, BORON COMPLEX/CN
E7           2     BENZENE, 1,2-DIMETHYL-, BORON-IRON-TUNGSTEN COMPLEX/CN
E8           1     BENZENE, 1,2-DIMETHYL-, BORON-MOLYBDENUM-TUNGSTEN COMP
                   LEX/CN
E9           3     BENZENE, 1,2-DIMETHYL-, BORON-RHODIUM COMPLEX/CN
E10          1     BENZENE, 1,2-DIMETHYL-, CADMIUM COMPLEX/CN
E11          1     BENZENE, 1,2-DIMETHYL-, CESIUM COMPLEX/CN


=> e "uridine, 5'-amino"/cn

E1           1     URIDINE, 5'-ACETATE CYCLIC 2',3'-PHOSPHATE/CN
E2           1     URIDINE, 5'-ACRYLATE/CN
E3           0 --> URIDINE, 5'-AMINO/CN
E4           1     URIDINE, 5'-AMINO-2',3',5'-TRIDEOXY-3'-FLUORO-/CN
E5           1     URIDINE, 5'-AMINO-2',3',5'-TRIDEOXY-5-ETHYL-/CN
E6           1     URIDINE, 5'-AMINO-2',3'-DIDEHYDRO-2',3',5'-TRIDEOXY-/CN
E7           1     URIDINE, 5'-AMINO-2',3'-DIDEHYDRO-2',3',5'-TRIDEOXY-, 
                   MONOACETATE (SALT)/CN
E8           1     URIDINE, 5'-AMINO-2',5'-DIDEOXY-/CN
E9           1     URIDINE, 5'-AMINO-2',5'-DIDEOXY-, MERCURY COMPLEX/CN
E10          1     URIDINE, 5'-AMINO-2',5'-DIDEOXY-, PLATINUM COMPLEX/CN
E11          1     URIDINE, 5'-AMINO-2',5'-DIDEOXY-5-(TRIMETHYLSILYL)-/CN


=> e hexane, 2,2,5,5-tetramethyl-/cn

E1           1     HEXANE, 2,2,5,5-TETRAKIS(METHYLSELENO)-/CN
E2           1     HEXANE, 2,2,5,5-TETRAKIS(P-(2,3-EPOXYPROPOXY)PHENYL)-/CN
E3           1 --> HEXANE, 2,2,5,5-TETRAMETHYL-/CN
E4           1     HEXANE, 2,2,5,5-TETRAMETHYL-3,4-BIS(METHYLENE)-/CN
E5           1     HEXANE, 2,2,5,5-TETRAMETHYL-3,4-BIS(METHYLENE)-, IRON 
                   COMPLEX/CN
E6           1     HEXANE, 2,2,5,5-TETRAMETHYL-3,4-DIPHENYL-3,4-DI-1-PYRENYL-/CN
E7           1     HEXANE, 2,2,5,5-TETRAMETHYL-3-METHYLENE-4-(NITROMETHYL
                   ENE)-, RADICAL ION(1-)/CN
E8           1     HEXANE, 2,2,5,5-TETRAMETHYL-3-PHENYL-/CN
E9           1     HEXANE, 2,2,5-TRIMETHYL-/CN
E10          1     HEXANE, 2,2,5-TRIMETHYL-, LITHIUM COMPLEX/CN
E11          1     HEXANE, 2,2,5-TRIMETHYL-3,4-BIS(METHYLENE)-/CN

Note that names must be searched EXACTLY as they appear in the record, including punctuation, Greek letters, etc..
In general, when searching names with complex punctuation, it is best to put them in quotation marks.
```
=> s "benzo(a)thiophene"/cn
```
You can truncate names, just like an exact title search in MELVYL. This can be very effective in retrieving multiple variations on a compound.
Remember, however, that there is an upper limit on the number of terms retrievable with truncation.

Heading Parent Field (/hp)

You can search the "parent" compound part of a chemical name separately in the /hp field. This has the advantage of not requiring truncation to find a whole family of compounds.
Remember that the choice of a heading parent can be tricky. Benzeneamine is the heading parent for anilines, not Benzene. Benzoic acid is the heading parent for benzene with a carboxylic acid group attached. So which is the heading parent for 4-amino-benzoic acid? Answer: Benzoic acid, but there's no intuitive way to know that.
You can combine /hp searches with chemical name fragments using the (L) operator

An example of an EXPAND in the Heading Parent field:


=> e 1,3,4-oxadiazole/hp

E1        1199     1,3,4-OXADIAZOL-2-AMINE/HP
E2          47     1,3,4-OXADIAZOL-2-OL/HP
E3        4098 --> 1,3,4-OXADIAZOLE/HP
E4        1680     1,3,4-OXADIAZOLE-2(3H)-THIONE/HP
E5           1     1,3,4-OXADIAZOLE-2(5H)-THIONE/HP
E6           3     1,3,4-OXADIAZOLE-2,3(2H)-DICARBOXYLIC ACID/HP
E7           1     1,3,4-OXADIAZOLE-2,5-D2/HP
E8           1     1,3,4-OXADIAZOLE-2,5-DIACETIC ACID/HP
E9           1     1,3,4-OXADIAZOLE-2,5-DIACETONITRILE/HP
E10         51     1,3,4-OXADIAZOLE-2,5-DIAMINE/HP
E11          1     1,3,4-OXADIAZOLE-2,5-DIBUTANOIC ACID/HP

Chemical Name Segment Field (/cns)

This field lists only the "basic" name segments. Its advantage is that you can use left-hand truncation in this field.
This can be very powerful for families of chemicals with similar roots in their names, e.g.
```
=> s ?cillin?/cns

=> s ?sterone?/cns

=> s ?porph?/cns
```

Chemical Formula Searching

SciFinder lets you search for exact chemical formulas, but the underlying Messenger command language has more powerful features, often very useful for the non-organic chemist.
The REGISTRY file makes chemical formula information searchable in a number of different ways.
Remember that the rules of the print Chemical Substance Index for molecular formulas can affect searches here too. Example: Copper (II) nitrate is Cu.2HNO3
The main formula field is Molecular Formula (/mf)
Formulas are written in Hill order, just as in the printed Molecular Formula index.
Remember that many substances can have the same MF, so MF searches are usually combined with some other search terms (e.g. name fragments) to narrow the search.
Multicomponent substances are written with a period separating the component formulas (see copper nitrate above.)
Polymers are written as, for example, (C2H4)x or (C2H4)n, depending on the case.
Copolymers combine the above, e.g. (C4H6)x.(C8H8)x for styrene-butadiene copolymers.

Element Symbol (/els)

You can search for the presence of a particular element (in any quantity) by searching the element symbol in the /els field. e.g.
```
=> s si/els and n/els
```
You can search for a number of atoms of a given element by searching the number followed by a slash and the symbol for the desired element, e.g.
```
=> s 6/c and 6/o
```
This is a numeric field, so it is range-searchable, e.g. si>=2.
You may search for all halogens (with X) or all metals (with M).

Atom Count (/atc) and Element Count (/elc)

With these two fields, you can specify the total number of atoms in the molecular formula (/atc) and the total number of different elements (/elc)
Both are numerically searchable.
These can get tricky with multi-component substances.

Element Formula (/elf)

This field allows you to search for a particular combination of elements without specifying the stoichiometry.
This can be very handy for substances like alloys or ceramic superconductors, where a given family of compounds of interest may have widely varying ratios of constituent elements and fractional stoichiometries
```
=> s ba cu la o/elf
```

Alloys in the Registry File

Since alloys are usually described by percentages rather than conventional molecular formulas, CAS has created special displays and search fields to accomodate them. While this information is displayble in SciFinder, STN and STNWeb allow you to search the information.
Note that CAS did not assign Registry Numbers to alloys before 1972.

Alloy Record Example


RN	39391-70-3
CN	Nickel alloy, base, Ni 75, Cr 20, Ti
		2.5, Al 1.5, Y2O3 1.3 (IN 853) (9CI)
CN	IN 853
CN	Inconel MA753
CN	MA753
DR	58719-29-2, 54259-74-4
MF	Al . Cr . Ni . O3 Y2 . Ti
CI	AYS
LC	CA, IFICDB, IFIPAT, IFIUDB
STE	8:AY, IN*853

Component		Component		Component
			Percent			Registry #
===============================================
Ni			75			7440-02-0
Cr			20			7440-47-3
Ti			 2.5			7440-32-6
Al			 1.5			7429-90-5
Y2O3			 1.3			1214-36-9

Alloys may have ranges instead of fixed percentages of composition.
Alloy Search Fields
- Material Composition Field (/mac)
  Range searchable, numeric field for percent composition. Examples
```
=> s (iron 70-74 and cr>10)/mac

 
```
- Relative Composition Field (/rc)
  Allows searching by component in decreasing order of percentage.
```
=> s fe.cr.ni.?/rc
```

Nucleic Acid and Peptide Sequences

All registered sequences of four or more nucleic acids or amino acids have displayble sequence information.
Amino acids may be written as either one-letter (ACDEF) or three-letter ('Ala-Cys-Asp-Glu-Phe') codes.
Nucleic acids use the familiar one-letter codes (A,C,G,T,U).
In the commercial version of STN online, you may search sequence information in two ways:
Sequence matching
- Available in STN online or STN on the Web.
- You may search for either exact sequence matches (/sqep, /sqen) or subsequence matches (/sqsp, /sqsn).
- You may specify gaps, wildcards, and repeated sequences, or whether a fragment must be located at the beginning or end of a chain.
- The SEQLINK command lets you work from a Registry number for a sequence to find other substances with the same sequence. This is the underlying command for the "Additional related references" option in SciFinder .
- For more information on this kind of biosequence searching, see Protein Sequence Searching, a reference guide in Adobe PDF format at http://www.cas.org/ASSETS/36138E32596F4965BE85D83F39DD88B9/protseq.pdf, and Nucleic Acid Sequence Searching at http://www.cas.org/ASSETS/4CE1649F453A44E78DC4763702375D92/nucleic.pdf.
BLAST similarity searching
- Available through STN on the Web, STN online using STN Express v. 6.01 or higher, or in the commercial version of SciFinder client.
- BLAST stands for the Basic Local Alignment S Tool. It is an algorithm developed at the National Center for Biotechnology Information (NCBI). It enables you to find similar biosequences, and ranks them by the degree of alignment.
- For more information on the BLAST algorithm, see the NCBI BLAST website at http://www.ncbi.nlm.nih.gov/BLAST/. For more information on searching using BLAST in STN, see CAS Registry BLAST Similarity Searching via STN on the Web at http://www.cas.org/ASSETS/DC25158BF1454BCFA13CB76C56C7E387/blast.pdf, or in commercial SciFinder, see Explore by Nucleotide or Protein Sequence at http://www.cas.org/support/scifi/howto/blast.html.

Polymer Searching

You can search for polymers in SciFinder by doing a substructure search on the repeating unit of the polymer, but it can be difficult to limit your search to polymers. STN online allows the use of substance class terms, one of which identifies polymers, to limit searches.
Another problem in polymer searching is that polymers may be represented by the structural repeating unit or by the monomer(s) which combine to form the polymer. As a result, the identical chemical composition may be represented by several different Registry Numbers. In STN online and STNWeb, the POLYLINK command allows you to find all the Registry Number for such a "family" of polymers starting from the Registry Number for any one of them. At present, this does not work for all classes of polymers, but it does for several of the major classes. SciFinder lacks this capability at the present time. For more information on POLYLINK, see STNotes #30 Enhanced Polymer Searching in REGISTRY at http://www.cas.org/ASSETS/51E8CEEC64F34F10A5154195E1237FBF/stnotes30.pdf.
Another problem: Chemists interested in polymers frequently want to search for information on broad classes of polymers, such as polyesters.
However, the common structural features of most polymer classes are so common that they cannot be structure searched in the REGISTRY file.
To deal with this problem, CAS has added polymer class terms (/pct) to the Registry File records for polymers.
There is a wide range of polymer class terms, including vinyl polymers, polyacetylenes, polyesters, phenol resins, etc.
You may add the term "formed" to a search to specify that a feature is formed in the reaction of the monomers.
A list of the polymer class terms may be scanned by doing an EXPAND in the PCT field in the Registry file, but it is best to consult search aid Polymer Class Terms in the REGISTRY File at http://www.cas.org/ONLINE/UG/polymerclass.pdf for definitions of each class if you intend to use them extensively.

Additional Structure Searching Features

Most of the power of Messenger structure searching is available in SciFinder; however, there are some useful features which have not yet been implemented.
Structure Fragments as R groups: Where SciFinder only allows you to define R groups in terms of atoms, variables and shortcuts, Messenger allows you to draw a structure fragment and define that as a variable group. Moreover, these fragments can have other variable groups within them, allowing for extreme flexibility.
Subset Searching: The SciFinder Refine option for substances allows you to do a structure search on a set of answers to narrow down your search. Messenger allows you to combine name or molecular formula based searches with structure searches for a more powerful version. You can also use ring structure features to search for your starting set.

Other Files in the CAS Family:

MARPAT

The REGISTRY file only indexes specific chemical substances. However, many chemical patents use Markush generic structures to specify classes of compounds covered by the patent. REGISTRY file searching alone will not retrieve these generic structrues.
MARPAT provides structure searchable records for patents with Markush structures described. MARPATpreviews contains records which have not yet received full subject indexing.
Records found are the CA records for the patents (including patent family information, where available), with Markush structure diagrams displayable.
For more detail, see the MARPAT Summary Sheet at http://info.cas.org/ASSETS/F9BE20739ECE46939A47D3FB793FFC44/marpat.pdf or the MARPAT User Guide in PDF format at http://www.cas.org/ASSETS/D5DF848824894DD7AE0CBDC0A5C49122/marpatug.pdf.
MARPAT information is not available in SciFinder .

Chemical Data Files

STN carries a number of files with numeric and textual chemical data information available in STN Online and STNWeb. In these files you can search both by compound and by data.
```
=> s 65-75/mp and 240-250/bp
```
Messenger has a SET UNITS command which allows you to select SI, metric or English units for either search or display. Units may be searched with tolerance ranges.
Most of these files have CAS Registry Numbers, so they will be listed in the Registry Record of a substance, and may be easily searched by Registry crossover.
General Chemical Data
- BEILSTEIN (structure searchable)
- GMELIN (structure searchable)
- HODOC (CRC Handbook of Data on Organic Compounds)
- MRCK (Merck Index)
Crystal Structure Data
- ICSD - Inorganic Crystal Structure Database
Thermodynamic Data
- CHEMSAFE (flammability data)
- DETHERM
- DIPPR
Toxicity Data
- HSDB
- MSDS files
- RTECS
- MERCK

SPECINFO -- Spectra Online

SPECINFO is a spectral data file, containing 13C, 17O, 19F, 31P and 15N NMR spectra, as well as some IR and mass spectra.
The file is searchable by name, RN, numeric data and chemical structure and has programs for calculating spectra.
Peak tables and actual spectra are displayable (the latter w/ STN Express).

Patent Searching

SciFinder displays patent family (i.e. patents from different countries covering the same invention, as defined by the priority patent application) information for the patents in the CAPLUS database. However, the patent family information is not searchable. In the STN and STNWeb versions of CAPLUS the patent family information is searchable.
STN has a variety of patent index files, some devoted especially to partcular countries, some with worldwide coverage.
Of particular interest:
- WPIndex (World Patents Index) Broadest of the files. Noted for descriptive titles, abstracts.
- INPADOC (International Patent Documentation Center) No detailed indexing, but covers 71 patent granting organizations, and legal status data from 40 patent offices. INPADOC is the source of the patent family data appearing in the CAS databases.
- IFIPAT (aka CLAIMS) U.S. Patents, detailed indexing, including CAS Registry Numbers.
- USPATFULL Full-test US Patents, crossover links to equivalent CAS patent listings, CA subject indexing and RN's added for chemical patents.
- EPFULL Full-text European Patent Office Patents.
- Other databases for specific patent offices: PATDD, PATDPA, PATDPAFULL (Germany), JAPIO (Japan), PCTFULL (World Intellectual Property Organization), FRANCEPAT, FRFULL (France), KOREAPAT (Korea), RUSSIAPAT (Russia)
- DPCI (Derwent Patent Citation Index) - Allows tracing by citations in patents, both by inventors and examiners, both patents and other documents.
- DGENE (Derwent Geneseq) Sequence searchable file of protein and nucleic acid sequences from patents., USGENE (sequence searchable file from the USPTO back to 1982), PCTGEN Sequence searchable file of protein and nucleic acid sequences from WIPO patents.

Citation Searching

The cited reference information which is displayable in SciFinder (for documents indexed 1997-present) can be used to "Get Related" cited references and citing references. However, the data is directly searchable by cited reference, cited author, cited year, etc., in STN and STNWeb. Note that cited reference information from both journal articles and patents is available.
The online version of Science Citation Index (SciSearch) on STN also has some advantages over the Web of Science version.
Searching in general is more powerful, and you can more directly combine cited reference and traditional author/keyword searching.
You can do the equivalent of a Related Record search (similar to the Web of Science feature) by using the SELECT command. SELECT pulls information from records and turns it into a search query. So you can take a record, SELECT RE (cited reference), search those references, then SORT OCC (sort by occurence) to get a Related Record set. This technique can also be used in the CAPLUS database, but the more limited availability of cited reference data (1999-present) makes it less useful.
SELECT CIT lets you create a cited reference search term from an article record. You can search a topic, then find all the papers that cited that initial set. This lets you search an author, then find who has cited her without worrying about the "first author" problem.
The main advantage of the STN Online/STNWeb approach over the Cited Reference cross-referencing in Web of Science, is that you may do your initial search in other databases, e.g., CAPLUS, and then crossover to SCISEARCH for the citation search.
You can also use Analyze features (similar to those in SciFinder ) to analyze citation patterns in more powerful ways than you can in Web of Science.
Note that, at present, the online versions of Science Citation Index only go back to 1974, while the Web of Science version extends to 1945 at UCSB and to 1900 at some institutions.
DPCI (see above) allows citation searching of patents, and the CAPLUS SCISEARCH and DPCI may be searched simultaneously for comprehensive retrieval.

Other Files Useful to Chemists

Besides most of the files available through UCSB's Web subscriptions (MEDLINE, INSPEC, COMPENDEX, NTIS, the Cambridge Scientific abstracts databases), there are a few additional bibliographic databases of particular use to chemists.
- ANABSTR - Analytical Abstracts (see the Important Indexes and Abstracts in Science and Engineering list for details.)
- NAPRALERT - Natural Products Database: A database of natural products, comprehensive from 1975 to date, selective from 1650 to 1975. Searchable by Registry Number, chemical name, organism.
The Messenger command language allows you to search any combination of these files simultaneously, and to identify and/or remove duplicate records from your searches.
Besides the CASREACT database which is used for reaction searching in SciFinder, other reaction searchable files on STN Online and STNWeb include:- CHEMREACT, CHEMINFORMRRX, DJSMonline. These have different chronological coverage and different indexing policies, making them a useful complement to CASREACT.
There are also many files of pharamceutical information, chemical business news, etc.

This page created by Chuck Huber (huber@library.ucsb.edu).
Updated: 03/02/08 04:42:39

Lecture 18 Extra: Why You Still Need to Go Online:Specialized Techniques and Databases