Output from Detect MSI Status

Three outputs are produced by the Detect MSI Status tool:

The MSI report contains both combined and per loci information on stability and other descriptive statistics related to the selected detection and test methods. The summary section contains a table of the analyzed loci and whether the individual loci are stable or unstable (figure 8.6).

Image detect_msi_status_report
Figure 8.6: Summary section from the MSI Report for an MSI-high sample.

The following information can be found in the locus table:

If the reads count is low but the coverage is high, it could be an indication that the locus is highly unstable and only few reads are spanning the locus. Investigating the read mapping can help understanding the problems.

Figure 8.6 shows an example of the summary section from an MSI report. In this case, 8 out of 9 loci are evaluated as being unstable using the coverage ratio method, meaning that the overall assessment of the sample is MSI-high. The only stable locus is NR21(A)21, but its coverage ratio (0.54) is only slightly larger than the threshold (0.49). When the value is close to the stability threshold, there is a risk of false evaluation, and in this case, it is likely that the NR21(A)21 locus is actually unstable. For the two first loci (BAT40(T)37 and MONO-27(T)27), on the other hand, there is a large difference between the coverage ratio value and the stability threshold (0.00 vs. 0.76 and 0.08 vs. 0.75), indicating that the loci are highly unstable. It is recommended to study the length distribution plots in the report for a more detailed view on the stability.

The baseline cross-validation report contains a table where the MSI status is presented for each sample in the baseline sample set. The cross-validation analysis verifies whether the baseline and selected parameters are suitable. For this, the MSI status of each sample from the baseline sample set is tested against a baseline created using all other samples of the set. Ideally, it is expected that all samples will be detected as stable (MSS) with a very low proportion of unstable loci. If this is not the case, the parameters might need to be adjusted and/or one or more samples should be removed from the baseline. Note that the cross-validation analysis is dependent on the parameters used for detection (exactly as for a test sample) and therefore each cross-validation is only valid for the selected set of parameter values used in the cross-validation run.