Full-Text Searching on MGI Query Forms
This help document describes the MGI search engine that matches words and phrases and allows Boolean connectives between those searches for words and phrases.
- What is full-text searching?
- Which fields on MGI Query Forms support the full-text search?
- What database records are searched? What does the search return?
- Which Boolean operators are allowed? How do they work? Is there a default operator?
- How do I construct a query using multiple terms?
- What if a term contains commas?
- Does case matter?
- What are the rules for full-text searching?
- What is word stemming and how does it work?
- What is partial word matching and how does it work?
- How do I use quotation marks?
- How do I use parentheses?
- Are there any restrictions on full-text searching?
- Are there examples?
- What should I do if I get an error message?
- What happens if my query returns no results?
See also:
- MGI Glossary for unfamiliar terms (or for how MGI defines familiar terms) and acronyms
- General Guidelines for MGI Query Form Searches
- Using Wildcards on MGI Query Forms
What is a full-text searching?
In full-text searching, the system takes the term you enter, looks for that text in all database records of the appropriate type, and returns a list of all records containing the term in the search string.
Note: If you enter more than one term, you will get back records that contain all the terms. See Which Boolean operators are allowed? How do they work? Is there a default operator?.
TopWhich fields on MGI query forms support a full-text search?
The following MGI query forms support full-text searches:
- Genes and Markers Query Form (phenotype/disease field)
- Gene Expression Literature Search (Text field)
- Phenotypes, Alleles, and Disease Models Query Form
What database records are searched? What does the search return?
The system takes the text you enter in the phenotype/disease box on the query form and searches the following MGI database records:
- MP term names, synonyms, accession IDs and annotation notes (General and Background-Sensitive notes)
- Allele Additional Information Field
- human disease vocabulary terms, synonyms, and accession IDs.
The phenotype/disease box is located at the top of the query form in the Mouse phenotypes and models of human disease section (beneath Enter any combination of phenotype terms, disease terms or IDs).
The Human Disease Vocabulary has a flat structure and the Mammalian Phenotype (MP) Ontology is hierarchical. Thus:
- For a Human Disease Vocabulary entry, the search returns alleles annotated to that term and that term only.
- For an MP term entry, the search also returns an allele if the genotype for that allele is lower down in the MP hierarchy. For example, entering a high level term such as heart returns not just alleles annotated to heart, but all genotypes annotated to cardiac hypertrophy, dilated atria, and so on.
Which Boolean operators are allowed? How do they work? Is there a default operator?
Two operators are allowed: AND and OR. NOT is currently unsupported but may be added at a future time. By default, a space between terms is interpreted as an AND.
The format is:
term OPERATOR term ...
- If you enter term1 AND term2, the search locates all records containing both terms in any order.
- If you enter term1 OR term2, the search locates all records containing either term1 or term2 or both terms.
Examples:
Top
If you enter... the search tool returns... heart AND defect Entries for heart defect but not for heart by itself or for defect by itself. Both expressions must appear in any matching terms. heart OR defect Entries for either heart or defect. heart defect Entries for heart defect but not for heart by itself or for defect by itself. (Same results as heart and defect, above.)
How do I construct a query using multiple terms?
By default, the search tool interprets a space between terms as AND. Therefore, if you do not want results for something containing term1 and term2 and term3, be sure to enter or between terms: entering diabetes and Parkinsonism, for example, (probably) yields zero results because there are (currently in MGI) no diseases containing both terms.
What if a term contains commas?
Commas between terms (or within terms) are interpreted as ORs. Therefore, if the term itself contains a comma and you want to search on the exact phrase, enclose put the entire term in quotation marks (example: "cyclin B1, related sequence").
TopDoes case matter?
No. The search engine is not case sensitive. Terms and operators may be uppercase, lowercase, or any mixture of the two.
TopWhat are the rules for full-text searching?
- Enter one or more terms or multi-word terms (e.g., circulatory, immune system).
- Separate words or phrases with one of the operators: AND or OR. Example: liver and pancreas
- Add "whitespace" (space or tab) on either side of any operator. Example: embryogenesis AND defect
- Enclose text in quotation marks if you want the search to return only items containing that phrase; e.g., "endocrine system"
- Enclose terms in quotation marks if you want the search to return only exact matches on that term; e.g., if you want regulation and not related words such as regulatory, put quotes around it. See How do I use quotation marks?
What is word stemming and how does it work?
The MGI search engine stems (i.e., cuts off) at the word level to retrieve matches that are close yet not exact. For example, you can enter diabetic or diabetes and get the same results because the system truncates your entry to diabet.
TopWhat is partial word matching and how does it work?
The search engine performs partial matches on words: it looks for terms that begin, end, or are contained in other terms. For example, if you enter ear, the system not only returns results for ear, but also for early and heart.
Therefore, be sure to put a space in front of any term that begins, ends, or is contained in another term and enclose the term in quotation marks. For example, if you want results for eye, you must enter “ eye”. If you don't, the system stems eye to ey and matches ey any where in a word, thus returning results for e.g., kidney as well as eye.
How do I use quotation marks?
- Enclose anything that you do not want reduced to a stem in quotation marks (quotation marks "turn off" word stemming). For example, if you want results for cardiac but not cardiovascular, enclose the former in quotes to prevent a return of everything beginning with cardio.
On the other hand, if you want to look at every allele or phenotype involved in immunity, simply enter the stem, immun without the quotation marks. - If the term you are entering contains a comma, put the entire term in quotation marks to ensure that the search is on the exact phrase (example: “cyclin B1, related sequence”).
- Be sure that all quotation marks are balanced. There must be one pair before and after each search term (e.g., “abnormal morphology”).
- Use only double quotation marks (single quotation marks - i.e, ‘cardio’ - are invalid.
See Are there examples? for additional entries and results.
TopHow do I use parentheses?
Parentheses tell the search engine the order in which it should interpret your terms. Consider two examples:
a or b and c e.g., cord or column and spinalBy default, the engine interprets AND first, and so evaluates the first example as if you had put parentheses around b and c. The result is a return of anything containing spinal column, or cord, including accordance, record, and so on. The second example returns results for spinal cord or spinal column.
(a or b) and c e.g., (cord or column) and spinal
See Which Boolean operators are allowed? How do they work? Is there a default operator? for how the search engine interprets AND and OR.
TopAre there any restrictions on full-text searching?
- Wildcards are not allowed.
- The NOT operator is not allowed.
- If you do not enter an operator, the default is AND between terms.
- The system ignores variations in punctuation and spacing. The results for each of the following entries are the same.
heart, defect
heart defect
heart and defect
Are there examples?
See also examples (with results specific to the Human Disease Vocabulary Browser).
| If you enter... | you'll get results for... | Why? |
|---|---|---|
| membrane | membrane, membranous; transmembrane anything containing membran | Word stemming and partial word matching. |
| hearing | hear | Word stemming. |
| hearing | hearing, heard, hears, heart | Partial word matching. |
| "hearing" | hearing shearing | Full word match Partial word matching. |
| "ear" | Anything containing ear including ears, hear, heart, early | Partial word matching. |
| " ear" | ear, ears, early | Quotation mark and white space at beginning eliminates partial word matches on hear, heart. |
| " ear " | Only ear | Leading and trailing quotation marks and white spaces prevent partial word matching at either end; no returns for early, hear, ears, or heart. |
| spinal cord | spinal cord | Implicit AND operator. |
| "spinal cord" | spinal cord | Quotation marks. |
| spinal or cord | Anything containing either spinal or cord | OR operator. |
| spinal and cord | Anything containing both spinal and cord | AND operator. |
| spinal, cord | Anything containing both spinal and cord | AND operator. System ignores extraneous punctuation. |
| spinal and (cord or column) | Anything containing both spinal and cord or column | AND parentheses, OR. |
| spinal and (cord or column | Error | Opening parenthesis is present; closing parenthesis is missing. |
| "abnormal spinal cord morphology | Error | Missing quotation mark. |
| 'abnormal spinal cord morphology | abnormal spinal cord morphology results | System ignores extraneous punctuation. |
| Factor V Deficiency | results for Factor V, VI, VII, etc. | Partial match on V. |
| "Factor V Deficiency" | Factor V alone | Quotation marks turn off stemming. |
| cytochrome c, somatic (gene/marker symbol/name field) | cytochrome c or somatic (~43 matches) | Quotation marks needed if desired search is on entire term (searching on “cytochrome c, somatic” results in ~1 match). |
| superimposed aqueductal stenosis | One allele with an allele name, chromosome and category information, but no observed effects or allelic composition (genetic background) data | The Mammalian Phenotype (MP) Ontology is annotated at the genotype level. The allele appears because of a match in the Additional Notes Field but it has no genotypes that contribute to the query result. |
What should I do if I get an error message?
- Check the punctuation in your search string and make sure that all necessary quotation marks and parentheses are present and balanced.
- Check your spelling and the sections on partial word matching and word stemming again.
- Retry your query, using more (or fewer, more specific) parameters.
- Change the number next to Maximum Returned.
What happens if my query returns no results?
If your query returns no results, the resulting report:
- Summarizes your query parameters
- Indicates that there were 0 matches
- May list additional resources to help refine your query.
The probable causes are:
- There is something wrong with your selection from one of the browsers (i.e., did you copy and paste the entire term/ID?) See the following help documentation for clarification:
- There is an error in the text search string in the Phenotypes/Diseases box. For example, did you enclose multi-word terms in quotation marks? See examples for help.
- The allele nomenclature is incorrect (see Quick Guide to Nomenclature for Alleles and Mutations).
- The number in Maximum Returned is zero.
- The database does not contain your term yet.