About the soma location when using the nblast

[MQChen]

Hi,

I was following the script which in github (https://github.com/natverse/nat.examples/blob/master/10-kunst2019/03-NBLAST.R, https://github.com/jefferislab/NBLAST_figures/blob/master/R/ExampleNBLASTTraceToFC.R), and I noted that before nblast() the neurons data, need dotprops() it to point clouds; so I want to konw different neurons' soma location will have influence on the morphology cluster? Like: neuron A & B (in a large enought daaset), have the similar morphology but locate in different region, after nblast will in different cluster?

Thank you. Best, MQChen

[Greg Jefferis]

Dear MQChen,

Somehow we missed this query last December – apologies.

The current NBLAST implementation is very simple and it does not pay any special attention to soma location – it is just another point. There are various ways that one could try to get round this when the target / query neurons have a soma. For example it would be possible to make a separate probabilistic score for this and add it to the main one.

Best,

Greg.