Table 4.

Summary of candidate antibody comparisons.

Target protein	Reference antibodies	Total no. of candidate antibodies tested	No. of good candidates chosen by expert	No. of good candidates chosen by SACT	SACT false negatives	SACT false positives
Gephyrin	GAD	7	5	4	1	0
Homer	PSD95	4	1	1	0	0
IRSP53	PSD-95, VGluT1	12	2	2	0	0 (+1 unclear)
VGAT	GAD	17	4	4	0	0 (+2 unclear)
Collybistin	GAD	19	6	6	0	1
Bassoon 1st exp.	Synapsin	19	4	4	0	0 (+1 unclear)
Bassoon 2nd exp.	VGluT1, PSD-95	10	5	5	0	0

Multiple candidate antibodies against the target protein were ranked using measurements from the proposed SACT. The candidate antibodies were independently evaluated by visual inspection of the immunofluorescence images. For each target protein, the reference antibodies were chosen according to the known characteristics of synapses expressing the target protein. Without an objective and automated tool to quantify antibody performance, all antibody evaluations are subjective (as is the standard today). The framework proposed in this work is a first step toward having synaptic antibody quantification be a routine part of antibody evaluation.