MedMiner

• Organizing genes on a chip
• Exploring one gene (Gene Query)
• Exploring two genes (Gene/Gene Query)
• Exploring a gene and drug (Gene/Drug Query)
• Typical problems

GeneCards Filter Example

Question: What genes on the chip are related to apoptosis?

The GeneCards Filter can be used to organize genes on the chip according to pathway, function, etc. First, enter "human_oncochip" at the top of the page. Then select "gene" for the query type. Next, enter "apoptosis" into the entry box. The genes on the chip which have the word "apoptosis" in their GeneCards are highlighted in red. All of these genes are listed at the end of the page to make it easy to organize genes on the chip.

top of page

PubMed Filter Examples

Question: How do I find out more about the gene Caspase 3 with GenBank Accession number U13737?

First, enter "human_oncochip" in the topmost box and select "gene" for the query type. Next, enter "U13737" into the text entry box. After the GeneCards query, select the box next to "U13737" and press the "Filter Literature" button at the top of the page. Select a 60 day time limit at the top of the PubMed Filter page. Check "CASP3," "CPP32B," "Yama," "CPP32," and "apopain." For additional comma-separated search terms, type in "caspase 3,caspase-3." For terms to narrow the search, type in "leukemia."

The first link on the results page will download all of the complete abstracts from PubMed. Next, a summary is provided which shows how many relevant sentences were extracted. As a compromise to balance precision and recall, a sentence is considered to be relevant if it contains one of the gene synonyms and a keyword. An abstract is considered to be relevant if it contains at least one relevant sentence and also contains the word "leukemia."

To navigate the tables, identify a keyword of interest and then click on the link. The number after each keyword is the number of relevant sentences found containing that keyword. To view the complete abstract, click on the PMID link. To see more tables, click on "top of page" and select additional keywords. To view any potential "false negatives," click on the link provided and then choose a PMID from the list.

top of page

Question: How do I find out about possible relationships between p53 and mdm2?

First, enter "human_oncochip" in the topmost box and select "gene_gene" for the query type. Enter "p53" in the first box and "mdm2" in the second box, then press the "Gene/Gene Filter" button. Select "X54156 TP53 P53 tumor protein p53 (Li Fraumeni syndrome)" from the first table. Select "M92424 MDM2 mouse double minute 2, human homolog of, p53-binding protein" from the second table. Press the "Filter Literature" button at the top of the page.

Select a 60 day time limit at the top of the page. Check the gene synonyms "TP53" and "P53." Enter "Li-Fraumeni" manually as an additional search term. Check the "MDM2" in the second table. Press the "Submit Query" button.

The first link on the results page will download all of the complete abstracts from PubMed. Next, a summary is provided which shows how many relevant abstracts were extracted. As a compromise to balance precision and recall, a sentence is considered to be relevant if it contains one of the gene synonyms and a keyword. An abstract is considered to be relevant if it contains at least one relevant sentence and at least one of the synonyms for the other gene.

To navigate the tables, identify a keyword of interest and then click on the link. The number after each keyword is the number of relevant sentences found containing that keyword. To view the complete abstract, click on the PMID link. To see more tables, click on "top of page" and select additional keywords. To view any potential "false negatives," click on the link provided and then choose a PMID from the list.

top of page

Question: How do I find out about possible relationships between p53 and Paclitaxel?

First, enter "human_oncochip" in the topmost box and select "gene_drug" for the query type. Enter "Paclitaxel" in the first box and "p53" in the second box, then press the "Gene/Drug Filter" button. Select "125973 Paclitaxel" from the first table. Select "X54156 TP53 P53 tumor protein p53 (Li Fraumeni syndrome)" from the second table. Press the "Filter Literature" button at the top of the page.

Select a 60 day time limit at the top of the page. Check the drug synonym "Paclitaxel." Enter "taxol" manually as an additional search term. Check the "TP53" and "P53" in the second table. Press the "Submit Query" button.

The first link on the results page will download all of the complete abstracts from PubMed. Next, a summary is provided which shows how many relevant abstracts were extracted. As a compromise to balance precision and recall, a sentence is considered to be relevant if it contains one of the gene or drug synonyms and a keyword. An abstract is considered to be relevant if it contains at least one relevant sentence and at least one of the synonyms for the other entity.

To navigate the tables, identify a keyword of interest and then click on the link. The number after each keyword is the number of relevant sentences found containing that keyword. To view the complete abstract, click on the PMID link. To see more tables, click on "top of page" and select additional keywords. To view any potential "false negatives," click on the link provided and then choose a PMID from the list.

top of page

Typical Problems

Problem: The results page is blank.

A blank results page may be caused by:

1) The Boolean search term was too long. Try using fewer synonyms, then resend the query.

2) The Boolean search term contained unbalanced parentheses or brackets.

The long drug names, in particular, might cause this problem. We will continue to work on a better solution for this problem, but currently the best option is to try to manually enter a simpler term, then resend the query.

3) There were too many abstracts to process.

There is a limit on the number of abstracts that MedMiner can process. If this limit is exceeded, a blank page may be returned. Try using a shorter time limit, then resend the query.

Problem: Some of the links on the results page don't work.

In order for the links to work, the page must be fully loaded before any of them are selected. Resend the query and ensure that the entire page has been loaded before trying any link. Also, see (3) above.

For other problems, send email to us.

top of page