STR tiering

STRs are only tiered for affected members of a family.

When estimating repeat sizes, ExpansionHunter provides confidence intervals and an average for each allele (i.e., x-y and avg(x,y)). For increased sensitivity, the maximum value (i.e., y) of these estimations is taken for each allele and locus to assign tier.

Note

The average value is displayed in the Interpretation Portal, as that value is expected to be closer to the real allele size. On rare occasions, a Tier 1 or Tier 2 STR expansion (tiered due to higher confidence interval exceeding the relevant threshold) may have an average allele size below the threshold applied during tiering.

STR loci that are green in PanelApp for the panel(s) applied for analysis will be tiered. Two different ranges of thresholds are used when tiering:

Tier 1

The repeat-length for the locus is greater than or equal to the pathogenic threshold.

Tier 2

The repeat-length is greater than or equal to the threshold for normal alleles but less than the pathogenic threshold.

Unified tiering of STRs and small variants

For biallelic loci (e.g. FXN), affected individuals homozygous for a pathogenic expansion or compound heterozygous (STR and SNV) are also tiered. This is achieved using the same approach as is used for tiering SNVs and Indels using the incomplete penetrance pipeline for an autosomal recessive mode of inheritance.

Note

Due to a limitation of the ExpansionHunter algorithm, in some cases, biallelic expansions of the FXN expansion may be incorrectly detected as monoallelic expansions.

Special notes regarding FMR1

Full FMR1 expansions (>200 repeats) cannot be distinguished from pre-pathogenic expansions and therefore in majority of cases will be reported as pre-expansions in Tier 2.

Tiering of FMR1 was introduced in the NGIS Grace release. For the cases processed with the earlier version of tiering, that was not considering FMR1 despite it being green in PanelApp, a warning will be displayed in the Interpretation Portal notifying that FMR1 was not considered in tiering.

Note

Reanalysis of cases that were referred before the NGIS Grace release will not include FMR1 STR tiering as these expansions were not detected prior to Grace, and the reanalysis does not include mapping and variant detection approaches.

Special notes regarding NOP56

There is a known limitation of ExpansionHunter’s performance for NOP56, where on some occasions it silently skips genotype calling for this locus when there are no reliably mapped reads due to locus complexity. This leads to missing NOP56 genotypes and therefore the testing for NOP56 should be considered as not performed. Clinical Scientists may want to caveat reports appropriately if there is no results for NOP56 displayed in the Interpretation Portal when they would be expected (i.e. STR locus is green on the panel applied).