Sensors in Entrez

As mentioned above, new sensors in Entrez include the Citation Sensor, the Accession Sensor, the Hot Topic Sensor for the H1N1 influenza virus, and the new Gene Sensor for sequence databases.

The new Citation Sensor automatically returns results from the PubMed Citation Matcher when it detects a query resembling a literature citation in a PubMed Search. Citation queries often retrieve irrelevant results when entered as a general PubMed search. The Citation Matcher service, now available as a part of the PubMed Advanced interface, is designed specifically for matching literature citations with PubMed records:

www.ncbi.nlm.nih.gov/pubmed/advanced

The Citation Sensor makes the power of the Citation Matcher more widely available. A minimal citation query would normally include an author name and a publication year or a journal name and publication year. For example, a search with “Lander 2001 Nature” quickly finds the Nature publication on the human genome sequence (Initial sequencing and analysis of the human genome sequence) as one of three articles found by the Citation Sensor (Figure 1, top panel). In comparison, the direct PubMed search retrieves 14 records, 11 of which are not from the journal Nature.

Figure 1

Citation Sensor and Accession Sensor in PubMed. Top panel. A search with “Lander 2001 Nature” showing the Citation Sensor. The Citation Sensor shows a more relevant set of results including the paper reporting the human genome sequence. (more...)

The Accession Sensor in PubMed is designed to provide relevant results when a PubMed search contains a sequence accession number. While GenBank sequence accession numbers reported in PubMed articles will find the source publication when used directly as a PubMed query, many accessions have no corresponding publication. Derivative sequence records such as NCBI Reference Sequences are often not associated directly with any PubMed records. Also, in many cases the goal of searching with accession identifiers is to find the sequence record itself and not the publication. In all of the above situations the accession sensor is quite useful in providing relevant results.

The middle panel of Figure 1 shows the results obtained in PubMed searching with a GenBank accession for the human dopamine D2 receptor (DRD2) mRNA (X51362). The search retrieves two PubMed citations that reference the accession as expected. The citation sensor in this case provides a convenient means to directly retrieve the sequence record without performing a separate search or following a link from one of the publications. The bottom panel of Figure 1 shows the results obtained using the corresponding NCBI Reference Sequence accession identifier for the DRD2 mRNA (NM_000795). There are no results found in PubMed since the RefSeq identifier is not cited in any publications or included in the abstract. However, the accession sensor provides access directly to the correct sequence record.

Another kind of sensor, the Hot Topic sensor, now appears in PubMed in response to increased searches related to the recent H1N1 influenza outbreak. In its present form, the sensor appears at the top of the right hand discovery column when it detects search terms that indicate interest in the H1N1 influenza sequences, and provides a link to the specialized H1N1 Influenza page described in the May, 2009 NCBI News (Figure 2, top panel). The Hot Topic Sensor will be deployed in different formats in response to current events in order to provide easy access to topical results.

Figure 2

Hot Topic Sensor, PubMed Central ad in PubMed, and Gene Sensor in PubMed. Top panel. PubMed results for a search with “influenza A” showing the Hot Topic Sensor link to the Flu sequences at the top of the right-hand column (boxed in red) (more...)

The Gene Sensor that has been active in PubMed for several months is now in the protein and nucleotide databases. As in PubMed, the Gene Sensor is triggered by a gene symbol in a search. The older sequence database gene search feature remains active and will still return results from the gene database when the search does not trigger the Gene Sensor. The middle panel of Figure 2 shows the Gene Sensor triggered in the nucleotide database by a search with the mammalian gene symbol AFM. The sensor allows retrieval of relevant gene records with access to nucleotide and protein sequences while the direct nucleotide results contain large numbers of irrelevant matches. The gene search results triggered by a search with “afamin” shown in the bottom panel of Figure 2 also provide a better set of results than the direct nucleotide search.

Database Ads

Two new Database Ads for the full-text PubMed Central database appear in PubMed. A link appears in all PubMed search results (Figure 2, top panel) displaying all articles that are also available in PubMed Central. Another ad for PubMed Central appears in the Abstract Plus record view and links to articles also in PubMed Central that cite the current article (Figure 3, top panel). This not only provides rapid access to full-text articles, but also offers another mechanism to expand the search to potentially related articles. As PubMed Central continues to expand the number of citations, it may also provide a useful measure of the significance of a particular article.

Figure 3

PubMed Abstract Plus and Protein GenPept view showing database ads and analysis tools. Top panel. Abstract Plus for an article reporting the 3-D structure of influenza haemagglutinin. The record has an ad for the 35 PubMed Central articles that cite the (more...)

A Structure Ad now appears in both the PubMed and sequence database record views (Figure 3). This ad features a thumbnail image of 3-D molecular structures reported in the PubMed article or linked directly to the sequence record. The image is linked to the corresponding record in the structure database. From here the structure may be displayed and manipulated in NCBI’s Cn3D structure viewer. In the sequence databases, records for influenza, dengue viruses, SARS, and retroviruses like HIV now display an ad for the taxon-specific viral genome resources area of the NCBI Web site. An example of the ad is shown in the bottom panel of Figure 3 for an influenza virus sequence. The viral resources pages have collections of viral sequences, genotyping and other specialized tools that virus researchers may find more useful than those within the general Entrez.

Analysis Tools

Direct links to sequence analysis tools in sequence records provide a means to instantly generate sequence-specific reagents through Primer-BLAST and update the annotation on all nucleotide and protein records through the ability to perform a live BLAST or conserved domain database search (Figure 3, bottom panel). Up to 20% of NCBI BLAST searches use NCBI database identifiers or copy-pasted NCBI formatted sequences as queries; the direct link to BLAST now makes it much easier to perform BLAST searches with NCBI database records.

Summary

New Discovery components in the NCBI System – Sensors, Database Ads, and Analysis tools – make the Entrez system more powerful and easier to use by providing context sensitive results that traverse traditional database boundaries. These components not only make it possible to find relevant information in fewer steps but also help make more obvious unanticipated connections that are often essential to scientific discovery.

BioSystems

NCBI BioSystems is a new database designed to aggregate biosystems information from collaborating public databases. BioSystems is a centralized repository of data that connects the biosystem records with associated literature, molecular, and chemical data throughout the Entrez system and facilitates computation on biosystems data. The NCBI BioSystems database currently contains biological pathways from the KEGG and BioCyc databases and is designed to accommodate other types of biosystems. Detailed diagrams and annotations for individual biosystems are available on the Web sites of the source databases. Links to Biosystems are now available from records in the NCBI Gene, HomoloGene, OMIM, and Protein Clusters databases. For more information, please see the BioSystems homepage: www.ncbi.nlm.nih.gov/biosystems/

Genome Resources

NCBI’s Genome Resource pages provide a comprehensive guide for a specific organism including links to NCBI resources as well as outside groups and consortia. New genome resource pages are available for the Pea Aphid (Acyrthosiphon pisum) and goat (Capra hircus). Links can be found under the “Organism-Specific” section of the Genomic Biology page: www.ncbi.nlm.nih.gov/Genomes/.

Microbial Genomes

Twenty-one finished microbial genomes were released between April 30 and May 28. The original sequence data files submitted to GenBank/EMBL/DDBJ are available on the FTP site: ftp.ncbi.nih.gov/genbank/genomes/Bacteria/. The RefSeq provisional versions of these genomes are also available: ftp.ncbi.nih.gov/genomes/Bacteria/.

RefSeq

RefSeq Release 35 is now available via Entrez and FTP. This full release incorporates genomic, transcript, and protein data available as of May 4, 2009. It includes 10,993,891 records from 8,393 different species and strains. Changes since the previous release can be found in the release notes on the FTP site. The RefSeq website is: www.ncbi.nlm.nih.gov/RefSeq/. The FTP site is: ftp.ncbi.nlm.nih.gov/refseq/release.

dbSNP

Complete data for the dbSNP Human build 130 are available on the FTP site and for searching on the web. More detailed genome build information is available on the dbSNP page: www.ncbi.nlm.nih.gov/SNP/snp_summary.cgi.