TABLE OF CONTENTS:

The form can take standard gene names (Gene Symbol) and gene identifiers from RefSeq and Ensembl.

The general format for gene names is the following:

Gene -> Official Gene Symbol (not aliases)
RefSeq -> NM_nnnnnn
Ensembl -> ENSGnnnnnnn.

As already mentioned, searches can be performed by genomic coordinates as well, taking into account that user provides valid coordinates from GRCh37/hg19 human genome assembly. We remind that the maximum length allowed is 350kbp, due to speed and transfer limitations.

The coordinates must conform the following pattern:    chr{A}:{start}-{end}
For instance, here we select the segment between position 79,349,683 and 82,346,281 of chromosome 4:    chr4:79349683-82346281

Last input field of the panel allows user to change the size of the flanking regions. Those flanks correspond to a genomic segment upstream/downstream the selected genomic loci (either if you provide a gene name or genomic coords). Just consider that the limit is 20kbp, both upstream and dowstream.

User can provide two positive integer values, both greater than 0 and separated by the slash symbol, as shown here: {upstream}/{downstream} (i.e. 3/5).
The first number redefines the upstream region (this depends on gene strand), while the second provides the downstream region length. when user provides coordinates or master SNP only, not gene names or identifiers, the strand defaults to forward ("+"). When gene names or identifiers are provided by user, then the corresponding genomic annotation strand is retrieved along with the start and end coords. When this field is set to 0/0, the program will add a minimum of 1kbp to both ends, upstream and downstream, just to ensure proper spacing on the final drawings.

The Master SNP corresponds to a SNP for which the user already knows whether it has functional annotated evidences or that has been found statistically significant from another experiment. By providing this SNP the user can get additional linkage disequilibrium (LD) information with respect to all the SNPs mapped over the genomic region appearing on the final genomic selected region. Such LD is computed for each single SNP on that region and the master SNP. That way the user can also get additional SNP candidates that are linked to the one provided as master; this can assist in the design of further downstream experiments. In the results table, both HTML or PDF versions, this master SNP will be prefixed by "**".

There are two types of analyses that can be done with a master SNP:

A) Search a gene/RefSeq/Ensembl or by coordinates, together with this SNP. It is important to mention that the user must know that the given SNP is within the coordinates given for the genomic location.

B) The master SNP itself, by leaving the coordinates and gene fields empty. By default, the web tool adds then 2kbp to the SNP genomic position, both upstream and downstream. Moreover, a flanking region can be also summed up to those new coordinates. Again, in order to define another flanking length, gene and genomic locus form fields must remain empty, otherwise the program will consider that SNP was part of the genic locus instead of an independent feature.

By default, if no cell line is selected on the corresponding panel of the input form, the search is made over all cell lines, and those falling inside the genomic segment selected will considered and drawn. On the opposite, if only one cell line is selected, the marks present in the given genomic locus will be drawn as one thick ribbon, in fact representing only marks relevant for that cell line (see DDIT4 example results). In the form, those cell lines lacking annotated features for that marks track are highlighted in italics and a gray color, to make the cell lines present in each track more readable. Multiple cell lines selection (default) has the advantage that the resulting figure can give visual clues about the chromatin state at the region of the selected SNPs, for instance.

This track will filter data for cell lines for which regulatory information is available, corresponding to promoters, promoter flanking regions, and enhancers.

This track will filter data for cell lines that have information about different histone marks, two of them found "abundant" at promoter sites (H3K9ac, H3K4me3) and the other two at enhancer sites (H3K4me1, H3K27ac).

This form field shows the available tissues for which an eQTLs can be recorded in the database.

The web application integrates information from different databases, as shown in the below chart. Those resources are pre-processed into MySQL tables by Perl scripts. On the web applicaation the information gathered from the users' input form is retrieved by PHP scripts, which use the provided parameters to select and filter out specific data from the local database tables and then store into temporary files to help in the making of the image maps. Finally, those PHP scripts also integrate those images and supplementary files into the different results summaries (PNG, HTML, PDF and ZIP files).

The following table summarizes all the information about the sources of the different marks, binding sites, SNPs, and eQTLs integrated in the web application MySQL database. This includes used database versions and links to the original resources. Some of the links point to a whole folder (like Promoters, Promoter Flanking, ...), because we downloaded the files from that folder. Links to UCSC refer to its "Table Browser" utility, further details of the queries can be found in the corresponding subsections below. The eQTL info is dinamically retrieved from GTEX site; now that site requires that users must register to acces the data, thus we provide the link to a tarball synchronized with this version.

The information was retrieved from the UCSC Table browser utility from the UCSC database. “Regulation group” table of the “TFBS Conserved” and “Txn Fac ChIP V2” tracks were selected. We cover 257 TFs having a total of 9,411,620 binding sites over human genome hg19.

Orange ribbons correspond to TF binding sites on the final plots (see graphical legend), whereas turquoise ones define CTCF-binding sites; overlapping TFBS are remarked as grey ribbons. We thought to highlight CTCF over other TF because it can function as transcriptional activator, repressor, or as insulator protein, blocking in that case the communication between enhancers and promoters. CTCFs can also recruit other transcription factors while bound to chromatin domain limits. The three-dimensional organization of the eukaryotic genome dictates their function, and CTCF serves as one of the core architectural proteins that help to establish this organization. The mapping of CTCF-binding sites in diverse species has revealed that the genome is covered with CTCF-binding sites.

Next table shows each TF and the number of binding sites over the genome that are annotated for it.

Light green ribbons represent the length of genomic locus defined by the user on the first panel of the web form; the query segment is defined by a pair of coords delimiting the gene (from its standard gene symbol, RefSeq or Ensembl identifiers) or the coordinates provided by the user. Orange and red ribbons refer to the 5'–flanking and 3'–flanking respectively. Blue ribbons represent a gene identified inside the input query (if exons are defined in the gene structure then this is the fill color for introns). Finally, dark green boxes inside the gene span represent exons. Gene names and Ensembl identifiers were retrieved from HGNC. Exon information and RefSeq identifiers were retrieved from UCSC Table Browser, from the “Genes and Gene Predictions” table of the “RefSeq Genes” track. We include genic structure annotations for 18,918 gene loci.

Data was retrieved from Ensembl Regulatory Database, release 84. Pink ribbons show promoter flanking regions, light purple ones depict enhancers, and beige ribbons correspond to promoters. We have included 100,483 promoter flanking regions sites, 140,349 enhancer sites, and 20,954 promoter sites.

Epigenetics data was downloaded from ENCODE. Different histone marks were mapped in order to show which modifications could reaffirm the promoter and enhancer sites mapped from the Ensembl database as functional. The histone modifications track includes: H3K9ac and H3K4me3, both correlated with promoter regions; and H3K4me1 and H3K27ac, both correlated with enhancer sites. Briefly, the H3K27ac summarizes 1,126,338 sites; H3K4me1, 1,516,873; H3K4me3, 420,127; and H3K9ac, 1,270,829 sites.

The post-translational histone modifications, H3K27ac and H3K4me1 are found to be representative of active enhancers. When they are both found in the same region and simultaneously in that region there is a mark for an enhancer, this can be considered as an evidence for that region to be transcriptionally active. Moreover, if a SNP disrupts both histone marks and the enhancer, this is a strong indicator that this SNP could affect the normal functioning of such a region. The same can be said about the H3K9ac/H3K4me3 pair, yet they are representative of active promoters. The number of blue and yellow marks represents the number of cell lines having those marks. By default, when no cell line is selected on the input form, then the search is made over all cell lines, and those having the mark will be drawn. On the opposite, if only one cell line is selected, if the mark is present in the given loci then it will be drawn as one thicker band, in fact representing only that cell. By setting the default web form parameter to search in all cell lines, the corresponding blue and yellow ribbons can provide a hint on the relative abundance of those marks. Again, filtering by cell line in the histone track will end up showing the marks that the cell line has and the corresponding ribbons will be thicker just to facilitate its visualization.

Data was obtained from Ensembl. In this track "open chromatin" feature is highlighted in dark green, H3K27me3 in purple, and H3K9me3 in light blue. The latter histone mark is related to heterochromatin yet it contains much less marks annotated; whilst the former is related to inactivation signals of the chromatin. Thus, we have considered appropriate to combine them in the same track. In sumary, there are 938,294 open chromatin sites, 3,236,553 H3K27me3 sites, and 11,882 H3K9me3 sites uploaded in our database.

Data was retrieved from the 1000GENOMES project. The track shows the SNPs that were mapped onto the genomic coordinates, mainly providing their location and dbSNP identifier. If a “master SNP” was provided in the initial form, the SNP label also contains the computed linkage disequilibrium (LD), the values are always shown within parentheses after the SNP identifier. Those cases without LD will show a "0" symbol inside the parentheses following the SNP identifier. SNPs remarked in green are also eQTLs, they are genomic loci that contribute to variation in expression levels of mRNAs. The “master SNP” from the initial form field will be shown in red. The current database covers 9,477,952 SNPs with MAF > 0,1 and 1,984,754 eQTLs.

The graphical legend summarizes the different tracks in the same order they appear in the results figures, starting from the left as the most inner track. For each track, the colors are annotated along with the genomic features represented in the plot. The graphical legend is embed in the HTML and PDF versions of the graphical results produced by this web application.

The figure below represents the results obtained after searching the FCHSD1 gene. As it can be seen, there are three additional genes in the query region: RELL2, HDAC3 and ARAP3. Each gene has its exons (dark green) and introns (blue) drawn, and to be able to differentiate between overlapping exon region in this track each overlapping gene will occupy a higher subtrack region (RELL2 in this figure). Those genes that are also in that region, but are not overlapping, will be drrawn in the outermost part of this track (HDAC3, ARAP3 and FCHSD1).

Initial Query:

- Gene: AKT1
- Master SNP: rs1130214
- Flanking: 10/10
- Cells: ALL
- eQTL: ALL
- SNP Population: ALL

Features that can be appreciated in the figure below:

- SNPs line beneath the tracks facilitating the visualization of the SNP's possible relationship/effect on the different structures (in case it disrupts the other features).
- Master SNP is colored red (find it on the lower right corner of the figure).
- eQTLs are colored green and each SNP has in parenthesis its linkage desequilibrium (LD) with the master SNP, or 0 if it does not have LD.
- Genes are colored blue (introns) with dark green as exons on top. The name of the gene includes the strand direction within the parenheses ("+" meaning forward strand, "-" for reverse).
- 3'-flank is shown in red and 5' flank in orange. If the input is a gene name/Ensembl/RefSeq IDs and it is in reverse strand (-), the 5'-flank will always be on the right side of the figure, whilest if it is in the forward strand (+), the 5'-flank will appear on the left side; the selected region (in light green) always fall on the bottom half of the plot (just below the main AKT1 gene in this example). Selected regions from chromosome coordinates and direct SNP queries are set in forward strand and will have the 5'-flank on the left.
- Each 1kbp of the input has an orange tick indicating the current position (GRCh37/hg19).
Note from this figure zoom, that the pair H3K9ac and H3K4me3 are more abundant in promoters-rich regions, and the pair H3K27ac and H3K4me1 in enhancers or promoter-flanking regions (they include enhancers).

In the figure below, we used the same initial query as in the previous case, but we also filtered the Chromatin, Regulation and Histone Tracks by the A549 cell line. We can see that this cell line presents same marks as in the previous figure, but here they are thicker.