With the new version, variants that are adjacent are reported as one variant (one row in the table view). Previously, if e.g. two adjacent SNVs were detected in the same reads, they would be reported as two variants (two separate rows), linked together in a linkage group. Each linkage group was given a namber and this number put in a column with the header 'linkage' in the variant track table. This caused a lot of confusion and interpretation problems for our users. Although we realize that changing the behavior of the variant callers will create disturbance in the analysis pipelines of our users, we have decided that we cannot ignore the feedback coming from a range of users reporting problems when interpreting the linked variants.
The change has a few consequences:
- We have introduced a new type of variants: the MNV (multi-nucleotide variant) as described in Variant types to hold variants that would previously be linked SNVs.
- Since only adjacent variants are reported as one, two variants that fall exactly on the first and third base of a codon will not be reported as one. They will be reported as two separate variants. This means that when calculating amino acid changes, it is not possible to unambiguously annotate these two variants. Instead, each variant is marked if another variant is present which could potentially alter its protein translation (there is now an extra column named "Other variants within codon").
- Variants that were previously reported as linked will be automatically converted to one variant when filtered and annotated. In addition, you can download a special plugin that will convert the data. The plugin is called 'Convert Variant Tracks' and is available in the plugin manager (see Installing plugins). Note that it is not necessary to convert the data before using it for analysis - it will happen automatically.
Please note that previously, linked variants would get one set of attributes, e.g. one count. When these variants are split either by the automatic conversion when creating a new track, or by the dedicated conversion plugin, each of the variants will inherit the attributes from the linked variant. In some cases, these values will be different from the values that would be calculated if the variants are calculated from scratch with the new version. As an example, the counts could be different when calculated separately for each variant compared to the count for the combined variant.
If it is important to ensure correct reporting of values for variants that were previously linked, we recommend rerunning the variant detection in the new version.