Constrained Random Alteration of Network Edges (CRANE) is a computational method for sampling networks with fixed node strengths. This package also include CRANE integraction with ALPACA.
You can install the released version of CRANE from GitHub with:
library(devtools)
install_github("PadiLab/CRANE")Generating Ensembles of Gene Regulatory Networks to Assess Robustness
of Disease Modules
James T Lim, Chen Chen, Adam D Grant, Megha Padi
bioRxiv 2020.07.12.198747; doi:
https://doi.org/10.1101/2020.07.12.198747
url: https://www.biorxiv.org/content/10.1101/2020.07.12.198747v1
We have included network edge list from angiogenic ovarian cancer (ang) and non-angiogenic ovarian cancer (nonAng). The following are some example use:
library(CRANE)
#>
#>
head(ang)
#> TF Gene weight
#> 1 AHR A1CF -1.090067
#> 2 AR A1CF -1.296286
#> 3 ARID3A A1CF 3.347115
#> 4 ARNT A1CF 2.394743
#> 5 BRCA1 A1CF 3.084647
#> 6 CREB1 A1CF -1.384014
# Running CRANE for bipartite network
newElist=crane.bipartite(ang,alpha=0.3)
#> [1] "Converting Edgelist to Adj Matrix"
#> [1] "Applying Alpha = 0.3"
#> [1] "Constructing Iteratuvely Perturbed Network"
#> [1] 100
#> [1] "Sorting Nodes"
head(newElist)
#> from to weight
#> 1 AHR A1CF -0.7996633
#> 2 AR A1CF -0.5410527
#> 3 ARID3A A1CF 2.9311514
#> 4 ARNT A1CF 2.3026895
#> 5 BRCA1 A1CF 2.8180388
#> 6 CREB1 A1CF -1.0686323library(ALPACA)
library(CRANE)
# Create Input Matrix nonAng as the baseline.
input=cbind(nonAng,ang[,3])
# Run ALPACA
alp=alpaca(input,NULL,verbose = F)
#> [1] "Detecting communities in control network..."
#> Weights detected. Building condor object with weighted edges.
#> [1] "modularity of projected graph 0.130474098565802"
#> [1] "Q = 0.138287700459528"
#> [1] "Q = 0.138287700459528"
#> [1] "Computing differential modularity matrix..."
#> [1] "Computing differential modules..."
#> [1] "Merging 12401 communities"
#> [1] 1
#> [1] 2
#> [1] 3
#> [1] 4
#> [1] 5
#> [1] 6
#> [1] "Merging 73 communities"
#> [1] 1
#> [1] 2
#> [1] 3
#> [1] 4
#> [1] "Merging 16 communities"
#> [1] 1
#> [1] "Computing node scores..."
# Apply Crane
alpListObject=alpaca.crane(input, alp, isParallel = T)
#> [1] "Converting to Adjecency Matrix..."
#> [1] "Generating CRANE Networks, this may take a long time (upto 10-20 minutes for Networks with 19000+ nodes) ..."
#> [1] "Creating Test Distribution..."
#> [1] "Detecting communities in control network..."
#> Weights detected. Building condor object with weighted edges.
#> [1] "modularity of projected graph 0.130474098565802"
#> [1] "Q = 0.138287700459528"
#> [1] "Q = 0.138287700459528"
#> [1] "Comuting Differential Scores..."
#> [1] "Computing Differential Score 20 %"
#> [1] "Computing Differential Score 40 %"
#> [1] "Computing Differential Score 60 %"
#> [1] "Computing Differential Score 80 %"
#> [1] "Computing Differential Score 100 %"
# TF Results
head(alpListObject$TF)
#> Membership Differential Modularity Score Pvalue t-statistic
#> ARID3A 1 0.11078839 7.100957e-43 -139.69642
#> ARNT 1 0.08121685 1.000000e+00 98.68718
#> BRCA1 2 0.06560763 1.000000e+00 23.98381
#> ELF5 1 0.03373797 1.000000e+00 98.33036
#> ETS1 2 0.05777723 3.636443e-30 -50.68560
#> FEV 2 0.05693467 1.000000e+00 87.91174
# Gene Results
head(alpListObject$Gene)
#> Membership Differential Modularity Score Pvalue t-statistic
#> A1CF 4 0.0014013382 9.999974e-01 5.5703422
#> A2M 1 0.0007451022 6.637922e-06 -5.2329474
#> A4GALT 6 0.0055156253 1.529687e-16 -16.4454533
#> A4GNT 2 0.0003932609 4.630008e-02 -1.7392323
#> AAAS 1 0.0003320119 1.000000e+00 9.1197478
#> AACS 9 0.0027239872 5.605418e-01 0.1536929