The Analysis window
Clicking the ‘Analysis’ tab produces a blank window and an ‘Analysis Panel’ with the ability to select the definition used to divide the STs into groups of related STs. The default setting is the sharing of alleles at = 6 of the 7 loci with at least one other ST in the group (see previous).
Clicking the ‘Compute’ button produces a display of the eBURST groups. For each group, the group number is shown, and the total number of isolates and STs in the group, and the predicted primary founder, if one can be assigned. Below this, all the STs in the group are shown, with the number of isolates of each ST in the input data (frequency), the number of STs in the group that differ at a single locus (SLVs), those that differ at two loci (DLVs), and at three loci (TLVs), and those that are more distantly related (satellites). Some STs may not share alleles at six out of seven loci (or five out of seven loci using the more relaxed group definition) with any other STs in the dataset and are shown at the bottom of the tabular output, below all the groups, as ‘singleton STs’.
A new feature of eBURST v2 is the average distance of each ST to each other ST in the group, which is computed from the average number of differences in their allelic profiles. Thus, in a simple eBURST group (e.g. that shown in Table 1), where all STs are SLVs of the primary founder, the average distance of the primary founder will be 1.0 as all other STs differ from the founder at only a single locus. The ST with the minimum average distance to other STs is likely to be the primary founder and this measure provides another indicator of the likelihood of a ST being the primary founder.
Table 1. Output for an eBURST group.
Group 7: No. Isolates = 20 | No. STs = 9 | Predicted Founder = 354
Average ST Bootstrap
ST FREQ SLV DLV TLV SAT Distance Group Subgrp
354 12 8 0 0 0 1.0 100% 100%
7 1 2 6 0 0 1.75 0% 0%
339 1 2 6 0 0 1.75 0% 0%
326 1 2 6 0 0 1.75 0% 0%
774 1 2 6 0 0 1.75 0% 0%
310 1 1 7 0 0 1.87 0% 0%
359 1 1 7 0 0 1.87 0% 0%
412 1 1 7 0 0 1.87 0% 0%
528 1 1 7 0 0 1.87 0% 0%
The conditional bootstraps are calculated for each ST being the primary founder of the group and, with the default setting, are displayed for re-samplings from an initial dataset that contains only one example of each ST in the input data (ST Bootstrap in the above table). Bootstrap values for re-samplings that use all isolates in the input data (which will typically include multiple isolates of some STs) can be selected in the ‘Analysis panel’. Bootstrap support is only provided with the default group setting, where all isolates in the group are considered to be members of a single clonal complex.
In eBURST v2, bootstrap support is also shown for each ST being a subgroup founder. These conditional bootstraps evaluate the number of times in the resampled datasets each ST is associated with the required number of SLVs to be defined as a subgroup founder. The default value for a subgroup founder is that it has at least two descendent SLVs, but this value can be increased, with corresponding changes in the level of bootstrap support as a subgroup founder. Thus, bootstrap support for a ST as a subgroup founder may be very high with the default definition but will decrease as the number of STs that are required to define a subgroup increases.
The number of bootstrap re-samplings can be changed from the default value of 1000 in the ‘Analysis Panel’. Bootstraps take a few seconds to appear and the progress of the bootstrap calculations can be followed in the ‘Analysis Panel’.