In the second method, we labeled populations of excitatory neurons based on projection target by using herpes simplex virus 1 (HSV1), which is efficiently taken up by axons and transported retrogradely.
In the first method, we labeled a population of cortical neurons mainly fast-spiking interneurons by using adeno-associated viral (AAV) gene transfer in a transgenic mouse line [18] in which the expression of Cre recombinase was driven by the parvalbumin promoter.
To demonstrate another potential application of the differential labeling method, we identified N-labeled (presumably female-derived) proteins by computationally adjusting the SEQUEST search algorithm to assume 95% N incorporation into peptides.
Note that this is not a statistical method and we label it as Optrv, "rv" for random variable.
In another class of problems, when even small random noise is injected into our neutron-proton data, another method, which we label HY, is superior.' In this ease, LM, NR, and KKH are sometimes completely ineffective.
In the second randomization method we used, node label randomization, we permuted the node labels in species A and compared it with the corresponding real network in species B in the same way as described above.
It should be noted that this step is taken only for CC, but in our method we handle missing labels differently using active learning, which helps in quantifying a probability score of interaction for each missing case instead of assuming it to be negative.
Indeed, in a test analysis with a TFT array using the random priming labeling method, we failed to detect all of the previously detected candidates (data not shown).
The labeling method we chose has the best cross-validation accuracy on average.
Using this method, we added F-enriched labels to different proteins, and visualized them by NanoSIMS and fluorescence microscopy.
