Different mode of inheritance in 1 search

[epouilli]

Hi,
Would it be possible to make ONE saved search where you get as a result:

• de novo variants (all types)
• homozygous (inherited) variants
• hemizygous (inherited) variants
• compound heterozygous variants
• known (likely) pathogenic variants (ClinVar) -> I know this option is already there
• loci associated with incomplete penetrance (I already made an idea topic for this: Incomplete penetrance #4624)
• other high impact variants -> This option is also already there

So I think the main thing here is the possibility of the first 4 items. The inheritance filter.

And to make it even better: sort the variants according to matching HPO therms...

Since most of the causal variants are 1 of these above, it would be really great that there is an option to make this as a "fast" first search so the causal variant can be discovered fast.

[hanars]

If you choose the "Recessive" inheritance mode it will return homozygous recessive, hemizygous x-linked recessive, and compound het results in a single search. We do not support combining dominant and recessive searches into a single search, so you would need to do 2 searches but not 4. This is because, generally speaking, you should be using looser population frequency criteria for recessive over dominant variation. Additionally, there is often familial and phentoypic context that supports one mode of inheritance as being more likely than the other. We already provide 2 recommended "fast" searches to do as your first search in seqr - the "De-novo/Dominant Restrictive" and "Recessive Restrictive" searches.

Sorting according to HPO terms is not possible - there is no deterministic algorithm for associating a genetic variant with HPO terms. While we have experimented in the with tools that perform phenotype-based prioritization of variants and genes this is not something that lends itself well to sorting. Instead, if you have a HPO terms of interest that have known disease-gene associations we recommend you either create a gene list for them or use one of the PanelApp gene lists in seqr and filter your results to that gene list.