Step 8 : ADO Calculation

The Tapestri Pipeline allele dropout (ADO) calculation module uses the germline variants from the given panel. It uses the ClinVar annotations to ignore the pathogenic variants and dbSNP annotations to find which common germline variants to use. 

The algorithm used to calculate the ADO is as follows:

1. Find all the heterozygous (HET) variants in the .loom file using the following rules:
   1. Minimum HET percentage: 0.7
   2. Minimum HET Genotype Quality (GQ) median: 0.6
   3. Variant Allele Frequency (VAF): between 0.4 and 0.6
2. For each variant identified as HET, check if the variant is called HET due to an allele drop out using the following logic:
   1. For each HET variant:
      1. Check if the variant is in the ClinVar database. If yes, skip it because it is likely a pathogenic variant.
      2. If the variant exists in the common dbSNP database, meaning it is a germline variant:
         1. Calculate the frequency of the HET variant.
         2. If less than 100%, its genotype is likely affected by the allele dropout, so the ADO rate is calculated for each variant as:
            1. Genotyped cells = Cells with reference calls (REF) + Cells with homozygous calls (HOM) + Cells with heterozygous calls (HET)
            2. ADO = (REF + HOM) / Genotyped cells
3. The ADO rate per run is defined as the median ADO for all variants identified in the previous step.
4. For each run, Tapestri Pipeline v2 provides the following details in the Run Report:
   1. Median ADO rate
   2. ADO calculation summary that includes:
      1. Number of variants per sample used for the ADO calculation and the name for each
      2. ADO rate per variant
      3. Percentage of genotyped cells
      4. Amplicon to which the variant belongs
5. As these annotations are specific to humans, we ignore these for the non-human samples and include all the variants that are identified as HET in step 1.