The best-fit SEM model fitted well with the data-based correlation matrix, P value was more than 0.56, root mean square error of approximation (RMSEA) equaled to 0, goodness-of-fit index (GFI) was more than 0.99.
In order to assess the extent of cell-type specific occupation of GAS sites and infer their roles in gene regulation, we first identified all GAS motifs around the peak center of STAT binding sites (+/− 75 bp) using the MOODS algorithm [ 26] with the known STAT position frequency matrix (p value < 0.01, JASPAR matrix id; MA014431 from http://jaspar.cgb.ki.se/) [ 27].
The test in (1) performed on each single clique identifies 11 cliques with significantly different covariance matrices (p-value ≤ 0.05) in the two groups of patients.
Invasive capability was reduced from 188.3 ± 4.04 (untransfected controls) to 56 ± 4.58 cells invading the matrix (p-value = 0.004).
After 24 hours, an average of 36.3 ± 5.03 sensitive cells and 188.3 ± 4.04 resistant cells invaded the matrix (p-value = 0.003).
We observed that, on average, the eGFP variants generated using the high expression matrix are statistically better expressed than those variants generated using the control matrix (p-value = 6.1e-6) and better than the wild-type eGFP (p-value = 1.4e-6).
For each potential duplicon of the original matrix, the P value is defined as the fraction of artificial matrices for which maximum copy number is larger than that of the potential duplicon.
All matrices with a P value of 0.01 or lower and with at least 25% coverage of the genes in the hit list are reported.
For each circuit, the mean and standard deviation were computed based on the appearance in each of the 1000 randomized matrices, and the P value and Z score were estimated by assuming a standard normal distribution.
CDCOCA produces a matrix of p values for all possible associations in the data matrix which are then used to enrich for associations dependent on sample complexity.
