Some SnpEff effects should be ignored

[Nicklas Nordborg]

The ANN field that is created by SnpEff can be a large table with multiple rows and columns. The order of the rows is affected by the "Impact" (HIGH/MODERATE/LOW/MODIFIER) of the effect. This order is not always the most practical order when trying to extract information since a HIGH impact effect may contain less information than a MODERATE impact effect. For example, the E545A mutation in PIK3CA (chr3:179218304 A>C) is annotated with 7 rows protein_protein_contact having HIGH impact but there is no protein change information. On row 8 there is a MODERATE effect that also contains the p.Glu545Ala (=E545A) protein change.

There are similar variants with structural_interaction_variant effect.

So when indexing and displaying information I think we should ignore effects with less information in favor for those with more information.

[Nicklas Nordborg]

In 6383:

Fixes #1326: Some SnpEff effects should be ignored

[Nicklas Nordborg]

In 6384:

References #1323: Genotype targeted variants

Ignore some effects from the ANN annotation. (see #1326)