connector counting discrepancy between CATMAID and NAT

[Arthur Zhao]

A question on the connector counting: the `syn count` in CATMAID's connectivity widget is not consistent with the number queried in R via `neuron$connectors`.  Do they mean different things?

If i look at the numbers between two partner neurons (a specific row of the connectivity matrix  in CATMAID VS `catmaid_get_connectors_between` in R), then the values seem to agree. 

See red markers in the attached file

[Greg Jefferis]

Hi Arthur, thanks for the question. This looks like it’s really down to the catmaid data model, but I’m open to suggestions for documentation improvements.

Basically one function is returning *connectors* and the other is returning *connections*. For polyadic connections, there is only one presynaptic connector. But that connector will make multiple connections (and that is what is being returned in the catmaid table).

Two additional points.

1. For postsynapses, the connectors function returns the connector id in the other neuron.

2. The connections between function should only produce a row for complete connections where both sides are defined.

Best,

Greg.

[Arthur Zhao]

Hi Greg,

Thanks for the helpful answer!  In terms of documentation improvements, maybe add a note on this?  Another idea is to emphasize the distinction between 'connectors' and 'connections' .

A follow-up question on addition point #1 -- isn't there a unique connector ID for each connection? I thought both neurons' $connectors list would contain this connector.

Best,

arthur

[Greg Jefferis]

On Wednesday, May 22, 2019 at 1:08:37 AM UTC+1, Arthur Zhao wrote:

Thanks for the helpful answer!  In terms of documentation improvements, maybe add a note on this?  Another idea is to emphasize the distinction between 'connectors' and 'connections' .

I think this might imply either adding a vignette that covers this information or adding it to documentation of individual functions. Where do you think you might have expected to find more information about this in the docs https://jefferis.github.io/rcatmaid/reference/index.html#section-synaptic-connections

 

A follow-up question on addition point #1 -- isn't there a unique connector ID for each connection? I thought both neurons' $connectors list would contain this connector.

No there isn't. There is a unique connector id for each presynaptic connector. The (typically multiple) postsynaptic partners are then differentiated by their treenode ids. So, to emphasise, a unique connection between a pair of neurons is indexed by two ids (connector+treenode).

[Arthur Zhao]

On Wed, May 22, 2019 at 8:03 PM Greg Jefferis <jeff...@gmail.com> wrote:

On Wednesday, May 22, 2019 at 1:08:37 AM UTC+1, Arthur Zhao wrote:

Thanks for the helpful answer!  In terms of documentation improvements, maybe add a note on this?  Another idea is to emphasize the distinction between 'connectors' and 'connections' .

I think this might imply either adding a vignette that covers this information or adding it to documentation of individual functions. Where do you think you might have expected to find more information about this in the docs https://jefferis.github.io/rcatmaid/reference/index.html#section-synaptic-connections

The 'Value' section in the documentation was the first thing i checked, since it's a confusion about what the returned values mean. 

 

 

A follow-up question on addition point #1 -- isn't there a unique connector ID for each connection? I thought both neurons' $connectors list would contain this connector.

No there isn't. There is a unique connector id for each presynaptic connector. The (typically multiple) postsynaptic partners are then differentiated by their treenode ids. So, to emphasise, a unique connection between a pair of neurons is indexed by two ids (connector+treenode).

Got it!