When viewing the results of analyses at a single transcript looking at all linked phenotypes, each point in the plot represents the association between a phenotype and the pLoF burden genotype for the selected transcript.
The x-axis position is grouped and coloured by phenotype category (the current groupings are Genomics’ standard phenotype definitions). The position on the y-axis is the -log10 P-value of the association between burden genotype and the phenotype in question, multiplied by the direction of the association – so a point above the x = 0 line indicates that loss-of-function increases the phenotype, and a point below x = 0 indicates that loss-of-function decreases the phenotype. The threshold for significance is based on Bonferroni correction and is stated below the plot. Quantitative and binary traits are plotted separately.
On the left side of the plot, dataset sources can be toggled on or off. At launch this includes phenotypes reported by UK Biobank and a small number of bespoke phenotypes derived by Genomics.
In the example below (figure 4), pLoF variant burden in the MANE (Matched Annotation from NCBI and EBI) select transcript of SLC2A9 (ENST00000264784) is significantly associated with decreased levels of serum urate (consistent with its known role as a urate transporter). Note that up to 10 significant phenotypes are labelled.
Figure 4:
The plot is interactive and allows the user to make some changes. Hovering over any point will show the association statistics. This is true for any phenotype, regardless of whether it was significantly associated with pLoF burden. The labels for significant associations can be moved by clicking and dragging. Clicking and dragging elsewhere on the plot allows the user to zoom into a defined area. Double clicking will reset the plot to its default zoom. Finally, hovering the mouse over the plot will reveal a small menu in the top right corner that allows the plot to be downloaded as a PNG file, along with functions for navigating the plot.
If any phenotypes are significantly associated with pLoF, a forest plot will be generated that indicates the size and direction of the effect, as well as the sample size (N), pLoF allele frequency (AF) and number of pLoF alleles (AC - allele count). The underlying data can be downloaded as a CSV using the download button on the right side of the plot. The plot itself can be downloaded as a PNG or SVG file.
Figure 5: An example forest plot
If no phenotypes are significantly associated after correction for multiple testing, a message will be displayed under the plot, as shown below for the binary trait associations for ENST00000264784.
Previous: How to Search by Transcript ID?