Created a separate issue from #53 for specifically about clonal assignment.
@bcorrie @(Gur?) as mentioned in the SciTech meeting, we can consider having an iR+ default clonal assignment pipeline. I think the current assign-clones app can be used for that. It can be parameterized some more to be more flexible.
For actually assigning clones to the current data in the ADC, I think it's reasonable that we can provide assignments in 2-3 different ways if we want. The # of clones should be an order of magnitude smaller than the rearrangements, so having a few loaded into the ADC seems reasonable. For example, for B cells and DefineClones.py in assign-clones we can use 1) a pre-defined threshold, and 2) the computed threshold; as two possible choices. In a way, that would be interesting to see the differences between the two across a broad range of repertoires.
Once we decide on what we want, kicking off a bunch of analysis jobs is relatively straight forward, then (at least on VDJServer) those analysis results are saved and can be loaded into the ADC anytime later.
Created a separate issue from #53 for specifically about clonal assignment.
@bcorrie @(Gur?) as mentioned in the SciTech meeting, we can consider having an iR+ default clonal assignment pipeline. I think the current
assign-clones
app can be used for that. It can be parameterized some more to be more flexible.For actually assigning clones to the current data in the ADC, I think it's reasonable that we can provide assignments in 2-3 different ways if we want. The # of clones should be an order of magnitude smaller than the rearrangements, so having a few loaded into the ADC seems reasonable. For example, for B cells and DefineClones.py in
assign-clones
we can use 1) a pre-defined threshold, and 2) the computed threshold; as two possible choices. In a way, that would be interesting to see the differences between the two across a broad range of repertoires.Once we decide on what we want, kicking off a bunch of analysis jobs is relatively straight forward, then (at least on VDJServer) those analysis results are saved and can be loaded into the ADC anytime later.