2-D plot of yeast expression QTL mapping results

• Sorry, the graph does not perform properly in FireFox. We have tested them under IE/safari + adobe SVG viewer 3 (ASV3). We also used javascript engine from ASV3 for stable performance across platform.

• BACKGROUND

  • For each of 112 yeast segregants, 6195 genes expression traits were measured using cDNA microarray;
  • over 3000 SNP markers were genotyped (Brem et al 2002 Yvert et at 2003);
  • Multiple Interval Mapping was applied to this mapping population to obtain expression QTLs (eQTL);
  • About 5000 QTLs across the genome were declared for about 3000 expression traits.

• TERMINOLOGY

  • Open Reading Frames (ORF) in the eQTL intervals are called "pORF" in this application since they are close to DNA polymorphisms. They are listed below "mapped onto:" in SVG.
  • Those ORFs, whose expression levels were used as traits, are called "eORF" hereafter. They are called 'trait' in eQTL mapping.

• BASIC GRAPHIC SETUP

  • Genes are arranged onto chromosomes according to their physical position;
  • chromosomes are concatenated, from chromosome A to P, to form x-axis, which are indicated using the checkered background (white and mintcream);
  • each small horizontal bar is an eQTL;
  • y coordinate of a bar corresponds to the gene whose expression was used as trait in QTL mapping (i.e. eORF);
  • start and end position of a bar correspond to the physical boundary of eQTL on the chromosome;
  • Check this toy example

• BASIC OPERATION

  • Press ctrl and use mouse to select and zoom in of a region;
  • Press alt and use mouse to drag the picture;
  • Notice zoom options when right clicking. We can go back to the original view anytime by choosing "Original View" ;
  • Dynamic annotation information : Place mouse pointer on a QTL, eORF and pORF (and also non-ORF genes in the eQTL) names will show on the right side of the screen. The sidebar will be fixed at that position no matter how the picture is zoomed or moved;
  • All QTLs of a same eORF will be highlighted with red color when putting mouse pointer at any QTL of the eORF;
  • Database Connection : Each gene name can be clicked and you will be led to annotation for it at http://www.yeastgenome.org ( It is recommended to right-click the name and to open the webpage in a new window);
  • Searchable : In the lower left corner, there is a box where trait name can be entered in by keyboard. QTLs for that gene will be marked (a red dot in the middle of each QTL bar). It is possible to mark a list gene stored in an external file (see next section).

• ADVANCED OPERATION

  • Gene interaction network can be superimposed onto the 2D plot to inspect visually the correspondence between known biological mechanism and eQTL detected through statistical analysis.

  • Using biochemical pathway as an example. When a pathway does contribute to the formation of eQTLs, it is expected to find relation I: pORF is in a same pathway as its eORF, or relation II: genes in a same pathway have overlapped eQTLs. Similar argument can be applied to other kind of known biological networks.

  • Available patterns/annotation datasets in our example to find relation I include
    • finalNet (big dataset, slow),
    • complex GO-slim (big dataset, slow),
    • MIPS,
    • two Hybrid,
    • synthetic Lethal,
    • transcriptional Factor,
    • over-Presented GO,(this annotation is generated by wei, using termfinder package),
    • KEGG (big dataset, slow).
  • It can be slow! Some groups of annotation contain huge amount of information. For example, finalNet contains 145254 pairwise gene relationships. It will take minutes for the plot to compare them with all declared eQTLs.

  • All annotation information is stored outside of the 2D plot. The left sidebar will disappear unless annotation is accessible to the plot. This will happen when users only download the plot itself from the server. Check prepare plot section for more detail. Check download section to download the complete example.

• AN EXAMPLE OF OPERATION

  • After loading the SVG file, click   "click for more Pattern" and find the purple word "KEGG". Click it to mark all eORF-pORF pairs, which are in a same pathway according to KEGG, with purple dots. Place mouse cursor on such a dot, the eQTLs' information will be shown on the right sidebar. Notice in the right sidebar, the pORF, which is in the same pathway as the eORF, is now tagged with a purple K and is highlighted.

  • click clear All to allow new patterns to be imposed.

  • Finding relation II: type KEGG.MARK in the search box at the lower left corner. You want to first click the box before you can type words in it. QTLs of genes in a same KEGG path way will be marked out with a same (random) color. Place the mouse point on the dots, we can above the search box the pathway the eORF/pORF belongs to. It is easy to find in the middle portion of chromosome O, there are many genes from oxidative phosphorylation pathway are mapped to similar genomic location.

See it!

In the above example, trait genes and QTLs are arranged in the same way: their physical positions in the chromosomes. In the next example, however, traits will be arranged along the y axis according to their functional similarity. Trait genes in a similar KEGG pathway are assigned similar y coordinates. Notice in this setting, a trait gene and its eQTL will be duplicated if it appears in more than one pathways.

Another example!