Chapter 20Using the Map Viewer to Explore Genomes

Data Sources

Non-Sequence-based Maps. Sources of maps that are not based directly on sequence include published maps in genetic, radiation hybrid, cytogenetic, and ordinal coordinate systems (where ordinal refers to clone order). The primary sources of each map are described in the online help documentation of each genome-specific Map Viewer. We are indebted to the researchers who make their mapping results so freely available. When a new version of any map becomes available, the data are also updated in the appropriate NCBI database.

Sequence-based Maps. The sequence-based maps shown through Map Viewer can be supplied by external sources and/or supplied from features computed within NCBI. For example, when the annotated sequence for a complete genome is submitted to the sequence databases (GenBank/EMBL/DDBJ), a copy of the data may also be accessioned as Reference Sequences (RefSeqs; see Chapter 18). The gene, transcript, and other feature annotations of the submitted complete genome are processed for display in the Map Viewer. NCBI staff may then calculate and display the position of other types of features, such as marker position or points of variation, as separate maps (Table 1).

Table 1

Types of Map Viewer annotation provided by NCBI.

Some of the annotation of genomic sequence carried out by NCBI is included in the genomic reference sequences (NC, NT, and NW Accession number format); however, other annotation is represented only in the Map Viewer and in the associated reports (Table 1). This latter type of annotation is based on information in several NCBI databases (Table 2) and is particularly important for attaching biological information to sequence data. Links to these resources are provided in Map Viewer to provide further information about each annotated object. It should be noted, however, that although sequence features may be placed in a genomic context automatically, there are curation steps that affect the final displays. For example, for the human and mouse genomes, sequences defining genes and pseudogenes are reviewed by collaborators and NCBI staff and, whenever possible, used as the basis of RefSeq records (NG, NM, and NR Accession number format).

Table 2

NCBI data resources used in NCBI-generated annotation.

Feature annotation is computed primarily in two ways: (1) by alignment of the defining sequence to the genome; or (2) for sequence tagged sites (STSs), by e-PCR (1). In some genomes, gene placement is based primarily on the alignment of mRNA [Expressed Sequence Tags (ESTs) and cDNAs], but only when an encoded protein is predicted. In other cases, where transcription evidence is weaker, more weight is given to identification of protein-coding regions. Gene identification is also constrained in that a known gene cannot be placed more than once in a haplotype (except for pseuodo-autosomal regions) or on an incorrect chromosome. Thus, if any reference haplotype retains inappropriately redundant sequence that encodes a gene, only one copy will be annotated as that gene. Others will be assigned interim IDs (see Chapter 14). Some ab initio methods may also be used for gene prediction. The predicted genes, as well as the mRNAs, are supplied as separate maps (gene, RNA, or GenomeScan maps).

In some cases, the position of these features may suggest the location of other genomic regions of interest. For example, the position of STS markers can help define the position of phenotypes such as quantitative trait loci (QTL). Although the best annotation of a gene or region is always through annotation by an expert researcher, automated annotation of genomes and comparison to that provided by experts can provide significant useful information. Experts interested in analyzing or assisting with genome annotation should contact us at vog.hin.mln.ibcn@ofni.

Relationships among Coordinate Systems

In addition to supporting the display of multiple maps in the same coordinate system (e.g., multiple sequence-based maps), Map Viewer also displays maps in different coordinate systems by calculating the correspondances among them (e.g., sequence to genetic). This is accomplished by: (a) identifying features that have been placed on maps in different coordinate systems; and (b) using general conversion factors. In the first case, placement of STSs on the genome is critical for the integration of sequence data with other, non-sequence-based maps, such as genetic and RH maps. The integration of cytogenetic data with sequence data is achieved through alignment of sequence from clones that have been placed cytogentically, such as the human fluorescence in situ hybridization (FISH)-mapped clones from the Bacterial Artificial Chromosome (BAC) Resource Consortium (2). The integration of non-sequence-based maps with the sequence provides a powerful mechanism to access portions of sequence on the basis of marker or cytogenetic data. Many features, such as Single Nucleotide Polymorphisms (SNPs), ESTs, mRNAs, whole genome shotgun reads, and clones can be placed on the genome assembly by using standard DNA sequence alignment methods such as BLAST.

The identification of known genes within the genome assembly provides critical landmarks and functional context to the sequence data, which in turn makes it easier to traverse to other rich sources of gene and protein information, including publications, OMIM, RefSeq, Conserved Domain Database (CDD), and LocusLink.

The power of calculating correspondances between coordinate systems may be more apparent when considering a common application of Map Viewer, i.e., identifying candidate genes within a region defined by genetic markers. When markers are palced on both genetic and sequence maps, it is then possible to use the gene-related maps (gene, UniGene/EST, or ab initio predictions) to identify possible genes of interest. For more details on how to do this, see the Map Viewer Exercises in Chapter 24.

A Work in Progress

For many genomes, identifying and positioning chromosomes and genes within sequence blocks is an ongoing process. In those cases, the Map Viewer can be used to evaluate the evidence that supports the current representation of the sequence and visualize possible conflicts. Inconsistencies in map order or in the placement of any object can be seen in the Map Viewer; this is assisted in some cases by the use of color coding (Figures 1 and 2).

Figure 1

Evaluation of a chromosome sequence (STS) map. Potential inconsistencies in the order or orientation of sequence blocks can be investigated by displaying a genetic map (Marshfield), radiation hybrid map (TNG), and sequence map (STS) together and checking (more...)

Figure 2

Evaluation of gene localization and annotation. A comparison of cDNA alignments (UniGene, RNA) and gene predictions (GenomeScan) to the genomic contig annotation can be achieved by displaying three maps simultaneously. The genomic contig (NT_024981.9) (more...)

For some genomes, the color-coded contig map displays whether the annotation is based on sequence assembled from draft or finished clones (blue, finished; green, whole genome shotgun; orange, draft). This is helpful when evaluating the level of confidence in the completeness of the annotation of a gene and/or its coding region.

Map Viewer also uses color coding or diagrams to represent the level of confidence in the placement of any mapped object. For example, SNPs or STSs that are placed at more that one position in a given map are noted by color (yellow) in the detailed labels (Figure 3a). Annotated genes are shown in different colors, based on the source and level of confidence in the annotation or the model (Figure 3b).

Figure 3

Representation of ambiguity. (a) The marker D1S2894 is found on several maps. Note that for the first map (STS), the circle is diagonally split with two colors. The diagonal means that the marker has been placed more than once; the two colors mean that (more...)

Frequency of Updates

Although maps provided from external sources are updated when new data are available, the maps dependent on NCBI's annotation process are updated periodically in versions called “builds”. Thus, mRNA or other supporting evidence that becomes available after the data “freeze” date for one build will not be incorporated into the display until the next build. However, some of the supporting databases linked from the Map Viewer may have more updated information. For example, UniSTS may provide more recent e-PCR results, or LocusLink may show a newer name or additional sequence data. dbSNP may make major data releases between builds; in this case, the variation map is updated.

Links from NCBI Resources

Many NCBI databases are now integrated into Map Viewer (Table 2); therefore, database records are often linked to Map Viewer displays. If a sequence in the public databases was released before the date of the current Map Viewer data freeze, then the position of this sequence may be displayed within Map Viewer. For example, Entrez Nucleotide, UniGene, UniSTS, and LocusLink records for sequences annotated on the human genome provide links directly to the appropriate region of the genome via links called Map Viewer (in the Links menu), Nucleotide, Map View, or mv links, respectively (Figure 4). It should be noted that such links are only precomputed if at least 50% of the sequence aligns with an identity of greater than 90%.

Figure 4

Connecting to Map Viewer from Entrez Nucleotide, using the Links menu in Entrez Nucleotide to connect from a record to Map Viewer.

Genome-specific resource pages also support queries via chromosome diagrams (Figure 5).

Figure 5

Example of a genome-specific resource page supporting queries to Map Viewer. Note that there are two ways to connect: (1) by selecting Maps from the pull-down menu in the gray bar at the top of the page and entering a query term in the Search box; or (more...)

Sequence Similarity Searches

Genome-specific BLAST pages that restrict a search to a specific genome are provided for several organisms and allow the results of the search to be displayed in a genomic context (provided by Map Viewer). Genome-specific BLAST searches can be accessed from the BLAST homepage, the Map Viewer pages of individual organisms (e.g., human, mouse), and the genome-specific resource pages of individual organisms. If the reference genome (the default) is selected as the database to be searched, the Genome View button (Figure 6) will appear on the BLAST results display page.

Figure 6

Accessing the Map Viewer display from a genome-specific BLAST results page. Selecting the Genome View button shows all of the BLAST hits on the genome.

Direct Query

Simple Searches

When already at a genome-specific Map Viewer page, any combination of query terms can be entered into a Map Viewer Search for box (Figure 7). Boolean operators (AND, OR, and NOT) and the use of * as a wild card (applied to the right of any term) are supported. The Search for and Help document hyperlinks provide current details about query options. An advanced search is available for some genomes.

Figure 7

Representative of species-specific homepage. Note the links to the help documentation and related resources. Also note the check boxes to use the advanced query page and/or to display objects calculated to have links to any object returned by the query. (more...)

Queries may include any unique identifier for a database record, e.g., a sequence Accession number or OMIM (MIM) number, or a text term or phrase, e.g., a gene symbol (BRCA2) or descriptor (p53-binding), or disease name (lung cancer). The Boolean AND operator is used automatically if multiple terms are entered. Therefore, a query for “fanconi anemia” will automatically be interpreted as “fanconi AND anemia”. The wildcard operator (*) provides a convenient mechanism to retrieve genes that share a common symbol or name, as is often found for gene families. For example, a query for ABC* will return matches to the ATP-binding cassette superfamily.

The advanced query page, accessed by checking the Advanced search box, provides additional options to refine a query. These additional options, which may vary from genome to genome, are useful for restricting queries to a particular search field or map type. The advanced query page also includes predefined search options to restrict the search to data with certain properties, e.g., to only find genes associated with a known disease or with sequence variation (SNPs). Additional refinements to queries against the variation map can also be made, for example, to search for variation markers known to be in a gene or coding region.

The same options for wild cards and Boolean operators for your query term(s) apply when starting at the Map Viewer homepage. At present, however, you must select a genome to which to restrict your search. An option to query across multiple genomes is under development.

Map Viewer Summary Results

The results from a query are displayed both graphically and in a summary table (Figure 8). When the query is executed by BLAST, the graphical view is color-coded according to the BLAST score, and the table summarizes the scores and the RefSeq Accession numbers that have matches. Clicking on the RefSeq Accession number (i.e., those beginning with NT_ or NW_) displays that BLAST result in the Map Viewer.

Figure 8

Map Viewer query results. (a) Note that an OMIM (MIM) number was entered as a query, and the Show linked entries box was checked. (b) The red tick marks next to the chromosome diagram indicate where the results appear to be placed on the chromosome. (more...)

Viewing the Maps

The Graphical Display

Text or Position Queries

General information on the chromosome being viewed is summarized at the top of the map page: the species and chromosome currently being viewed, the query term, and the name of the focal map, termed the Master Map (Figure 9c).

Figure 9

Representative map display. (a) Search and find options. (b) Compress Map allows you to select whether to display labels on maps other than the rightmost one. The Region Shown: and zoom graphic allow you to reset the range displayed either explicitly (more...)

The summary also includes the following statistics concerning the number of objects on the Master Map, which are:

• the number of objects localized (positioned) on the chromosome
• the number of objects not localized but present on the chromosome
• the number of objects localized in the region displayed (i.e., the number decreases as you zoom in)
• the number of objects for which text descriptions are shown (dependent on user-defined page length)

A thumbnail map on the left of the page provides a coarse indication of the region displayed; by default, this is a cytogenetic map, although the Master Map can be selected (Figure 9b).

Maps are displayed vertically, with the name of each map hyperlinked to a description of it (Figure 9d). Features displayed on the Master Map have brief descriptive labels; information on features on the non-Master Maps can be found by mousing over an object. The labels on the Master Map depend on the type of object and genome being explored but can provide: (a) links to resources defining the mapped element, some of which may not be at NCBI; (b) indicators of the confidence in the placement or naming or sequence in the region; (c) biological features of the element (for SNPs, this includes position in a gene or effects on the coding region); (d) direction of transcription for genes; and (e) links to tools to facilitate reviewing of the sequence (sv), downloading a subsequence of interest (seq), the mRNA alignments in a region (ev), homology maps (hm), or to create cDNA sequences in real time (mm). (See the section on Associated Tools for more information.).

Sequence (BLAST) Queries

The positions of BLAST hits are highlighted on the Contig map, and a text summary of the BLAST hit is provided with links to regional alignment reports. All of the options described previously for configuring your display are still available. Thus, it is possible to evaluate the sequence match by the location (possible intron/exon structure, percent identity) as well as to determine whether the matching genomic region contains all of the query sequence in the expected order. Adding other maps to the display using the Maps&Options window provides a powerful mechanism to determine how the query sequence corresponds to existing annotation, such as genes, gene predictions, STS markers, or SNPs. For more hints, see the tutorial section on querying the human genome by sequence.

Selecting the Region to Display

The Map Viewer provides zoom, navigation, and other map display controls. These can be found on the display page itself and in the Maps&Options window (Figure 10).

Figure 10

Representative Maps&Options window. The menus and boxes in this window allow definition of the range of the chromosome to display (Region Shown:), selection of the new map(s) to add to the display (Available Maps:), establishment of the order (more...)

As the resolution of a view is changed, the chromosome diagram is updated. The view automatically centers on a highlighted query term, or on the middle of the chromosome if browsing only. The chromosome view can be moved up and down or zoomed in and out. Zooming can be achieved in several ways: (1) by using the zoom control, located in the left column; (2) by providing a range or bounding markers in the Region Shown text boxes; or (3) by selecting part of the map to display a menu with predefined zoom levels. Most menu-based zooms should be carried out in two or three steps to avoid missing the region of interest. It is also possible to scroll or reposition the display by selecting recenter from the menu that pops up when you click on the chromosome diagram at the left or on a map or by clicking on the arrows at the top and bottom of a map.

Selecting the Maps (Tracks) to Display

Maps are categorized by the coordinate system as well as type of feature. The maps available for a genome can be seen by scrolling through the Maps menu in the Maps&Options window or in the genome-specific help documentation. For display and query purposes, different types of features annotated on sequence coordinates are treated as different maps. The maps in sequence coordinates are comparable because all of the sequence maps are based on the reference to a standard genome assembly. Thus, one can display the SNP map (at high zoom level) next to the Gene, UniGene, or GenomeScan map to ascertain the number and location of polymorphisms in a region.

Some basic map controls are available directly on the display including removal of a map from the display by clicking on the X over the map and moving a secondary map to the Master Map position by clicking on the arrow next to the map label.

The Maps&Options window provides advanced options to: (a) add a ruler to any map; (b) reset the page length to display more (or less) information; (c) define region to display by providing coordinates or marker name in Region Shown boxes (also available directly on the Map Viewer display); (d) display direct connections between maps by checking the Show connections box; (e) optionally view text in Verbose or Condensed mode by selecting the checkbox. These user-defined preferences will be maintained for additional queries on different regions or chromosomes, until reset.

There has been considerable effort to integrate data on the sequence-based maps with data from non-sequence-based maps. Map connections provide a unique and powerful mechanism to identify features in a relevant region of the sequence map when starting with information from a different coordinate system (see Relationships among Coordinate Systems).

The features that are available with Map Viewer are summarized in Box 1.

Box 1

Map Viewer-associated functions.

Data Access

The data displayed in Map Viewer are freely available. In addition to the view-specific reports, all of the data are available by FTP. README files document the content and format of each file. Genomic data are also available by chromosome; this includes genomic contigs (NT_ or NW_ Accession numbers) built from finished and unfinished sequence data. The contig data are available in various formats, including ASN.1, FASTA, GenBank, and GenPept. Also available in this directory are the RNAs (NM_, XM_ , and XR_ Accession numbers) and proteins (NP_, XP_).

Constructing URLs to Generate Specific Displays

Dynamic links to Map Viewer can be generated by constructing URLs with arguments that define the species, chromosome, range, types of maps (with or without units), display order, number of labels, query string, how to center a display around a query result, and the type of label for the display. The most current documentation is provided in the online help. The examples in Box 2, however, may illustrate the flexibility of the approach. Please note that the argument of the map in the URL is processed as an ordered list, with the order in the list controlling the left-to-right order in the display. Additional qualifiers control the display of a ruler and the range on the chromosome. If a query term is included as a part of the URL and that value cannot be identified on any of the maps in the list, that map will not be displayed.

Box 2

Examples of URL construction.

Implementation

Query terms are indexed for retrieval using the Entrez system. Thus, wild cards, Boolean operators, filters, and properties are managed as for other Entrez databases.

Each distinct object on the map is assigned a unique identifier that is specific to a particular build. Each object may have other secondary identifiers, such as IDs, in the sequence, Clone Repository, dbSNP, LocusLink, UniGene, or UniSTS databases. All descriptors are indexed as text. In addition, some are indexed by specific field values or by pre-identified properties, such as genes with associated diseases, SNPs with heterozygosity values in pre-defined ranges, or evidence type for genes. These field names or properties can be applied to restrict a query either in the Web-based query form or within a URL. The complete listings of current implementations for field qualifiers and properties are provided in the online help documentation.

Data for each map are retrieved for display from a relational database based on the IDs returned from the Entrez query. The database is used only to support display; it is refreshed with each NCBI build or update of any other map but not to track changes from build to build. Data from previous builds are archived at NCBI, but direct access is not currently supported.