Summaries of connectivity at the type level.

[dosumis]

Wondering whether this type of summary might be useful.

Queried for connections between neurons of types in the set

L2, L3, Mi1, TM2, Tm4, T4 & T5

upstream_type	weight	downstream_type	Robustness	downstream neurons
L2	21	Tm4	93.95	2.37
L2	104	Tm2	54.54	1.11
L3	18	Mi1	40.25	1.03
Mi1	16	T4	99.77	7.25
Mi1	13	Tm3	93.58	2.42
Tm2	13	T5	99.55	4.5
Tm3	9	T4	99.32	4.38
Tm3	6	Tm3	64.32	1.48
Tm4	8	T5	97.84	3.64

Weight: mean number of synapses per pairwise connection. Robustness: percentage of cells of the upstream type connected to cells of the downstream type - the number connected to a neuron of the downstream type/number of neurons of the upstream type for which we have connectomic data. Downstream neurons: mean number of neurons of the downstream type that each neuron of the upstream type connects to (dividing the number of pairwise connections by the number of upstream neurons.)

Visualisations - connectivity matrix - could use something fancier than a heat map to allow a couple of values in each square. e.g. see visualisations from transcriptomics of %expressing cells + levels using colors and bars

queries used:

MATCH (up:Class:Neuron)<-[:SUBCLASSOF*0..2]-(upsub:Class:Neuron)<-[:INSTANCEOF]-
(n1:Neuron:Individual:has_neuron_connectivity)-[st:synapsed_to]->(n2:Individual:Neuron:has_neuron_connectivity)-
[:INSTANCEOF]->(downsub:Class:Neuron)-[:SUBCLASSOF*0..2]->(down:Class) 
WHERE st.weight[0] >= 5 
AND up.short_form in ["FBbt_00003720", "FBbt_00003721", "FBbt_00003731", "FBbt_00003736", 
"FBbt_00003776", "FBbt_00003790", "FBbt_00003791", "FBbt_00003792"] 
AND down.short_form  in ["FBbt_00003720", "FBbt_00003721", "FBbt_00003731", "FBbt_00003736", 
"FBbt_00003776", "FBbt_00003790", "FBbt_00003791", "FBbt_00003792"]    
RETURN DISTINCT count(distinct n1) as upstream_count, count(distinct n2) as downstream_count, 
up.symbol[0] as upstream_type, round(avg(st.weight[0])) as weight, 
count(st) as pairwise_connections, down.symbol[0] as downstream_type
 order by pairwise_connections desc, weight desc

MATCH (up:Class:Neuron)<-[:SUBCLASSOF*0..2]-(upsub:Class:Neuron)<-[:INSTANCEOF]-
(n1:Neuron:Individual:has_neuron_connectivity) 
WHERE  up.short_form in ["FBbt_00003720", "FBbt_00003721", "FBbt_00003731", "FBbt_00003736", "FBbt_00003776", "FBbt_00003790", "FBbt_00003791", "FBbt_00003792"] 
RETURN DISTINCT up.symbol[0], count(n1) as upstream_count