Statistical and visual differentiation of subcellular imaging

BMC Bioinformatics

Table 1 Worst case p-values for subsets of Image set A

n (# images)	5	6	7	8	9	10	11	12	13	14	15	16	17	18	19	20
Worst p-value	.665	.541	.428	.611	.293	.249	.144	.245	.128	.103	.077	.045	.034	.055	.047	.017
Worst p-value (no outliers)	.337	.267	.134	.076	.117	.045	.080	.043	.029	.023	.012	.008	.007	.006	.005	.002

For each pair of the 10 classes of Image Set A, a subset of n images was randomly selected from each, and a p-value calculated for the null hypothesis of no difference. This was repeated 20 times for each pair. For a given n, the highest p-value over all pairs and their 20 repeats is shown in the table. P-values are shown for the cases of removing or not removing outlier images before random selection.

ISSN: 1471-2105