Using a clustering threshold of 49.2°, which was the average of the standard deviations of the six angles, 12 clusters were obtained.
Whole-brain analyses for the pain > no-pain contrast are reported using a cluster-forming threshold of p <.001 (uncorrected, cluster size > 10), with cluster-level family-wise error (FWE) correction.
The starting point of the technique is the definition of the output centroids using a clustering method compatible with most thresholding techniques in the literature.
For the whole brain comparisons, we used a cluster-level threshold corrected for multiple comparisons of P < 0.05 and an uncorrected voxelwise threshold of P < 0.001 (Fig. 2 and Supplementary Fig. 5, Table 1 and Supplementary Table 3).
In the case of cluster-level tests, we used a cluster-defining threshold of P < 0.001 (uncorrected).
The resulting clusters of differences in GM volume were corrected for multiple comparisons using random field theory with a cluster threshold of t >2.3 and a reliability of p < 0.05 for extend of the cluster.
Evaluation of the conformer model accuracy using RMSD is an intuitive and convenient choice, as the conformer model clustering in PubChem3D uses an RMSD value as a clustering threshold; however, in practice, PubChem3D primarily uses three measures in 3-D similarity comparison between molecules: shape-Tanimoto (ST), color-Tanimoto (CT), and combo-Tanimoto (ComboT) [37 40].
Therefore, an implemented cluster analysis was performed using a quality threshold algorithm (QT-clustering), in which we only considered genes with differences in transcript levels between the two genotypes at the three time points, i.e. 20 genes for time 0, 190 genes for time 2 h and 414 genes for time 24 h.
We use a minimum cluster size Cmin=5, a clustering threshold Z θ =0.3, and a distance threshold d θ =150 m.
