Trouble with WGCNA gene dendrogram
0
0
Entering edit mode
8 weeks ago
jms2520 ▴ 20

I am trying to plot my gene dendrogram while following the online tutorials for WGCNA. When using the function "table(bwnet$colors)" it shows that there should be 24 modules for my data. When I continue running the code for plotting the dendrogram (code below) I somehow keep getting a plot that looks like this. It seems to me that this is not consistent with there being 24 modules.  bwnet = blockwiseModules(datExpr0, maxBlockSize = 2000, power = 6, TOMType = "unsigned", minModuleSize = 30, reassignThreshold = 0, mergeCutHeight = 0.25, numericLabels = TRUE, saveTOMs = TRUE, saveTOMFileBase = "TOM-blockwise", verbose = 3) # open a graphics window sizeGrWindow(12, 9) # Convert labels to colors for plotting mergedColors = labels2colors(bwnet$colors)
# Plot the dendrogram and the module colors underneath
plotDendroAndColors(bwnet$dendrograms[[1]], mergedColors[bwnet$blockGenes[[1]]],
"Module colors",
dendroLabels = FALSE, hang = 0.03,
addGuide = TRUE, guideHang = 0.05)

table(bwnet$colors) ![> 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 1287 4001 3115 2498 1774 1706 1135 735 484 305 299 269 265 157 117 111 104 101 77 71 68 63 61 23 24 60 44]  ![enter image description here][1] Soft RNAseq WGCNA Thresholding • 691 views ADD COMMENT 0 Entering edit mode what is the output of table(bwnet$unmergedColors)?

0
Entering edit mode
table(bwnet$unmergedColors) 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 359 1273 718 1698 302 1041 533 424 781 579 159 108 106 92 89 365 366 302 279 164 123 90 79 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 79 52 37 225 215 178 174 161 156 119 116 87 80 74 73 72 62 58 51 741 416 363 283 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 36 34 948 260 162 136 113 102 631 434 396 110 88 74 242 194 191 157 136 100 88 85 82 69 70 71 72 73 58 45 35 35 33  ADD REPLY 1 Entering edit mode my guess is that you are plotting the dendrongram of the genes from block 1 bwnet$blockGenes[[1]] which happen to belong to grey (0), blue (2) and brown (1) modules.

By default, the maximum block size is 5000so the maximum number of genes you can plot in the dendrogram can't exceed that number

I am just guessing because I never used the function blockwiseModules

0
Entering edit mode

I decided to choose blockwiseModules because my RNA seq is analyzing 18000+ probes. So that would mean I may have to plot each block separately to get a visual? I was mostly concerned that something mathematically went wrong that I missed in the network construction because my dendrogram looked so different from the ones in the tutorials but perhaps it is because my dataset is so much larger and the maximum block size?

0
Entering edit mode

the code looks fine, and the dendrogram look like that because you are plotting only the genes from two modules that have opposite expression patterns = a gene dendrogram with only two main branches.

edit:

because my RNA seq is analyzing 18000+ probes.

How much RAM do you have?

0
Entering edit mode

I am working on a laptop with only 8 gigs

0
Entering edit mode

I am working on a laptop with only 8 gigs

Yeah, 8 gigs is not enough for 18k probes

0
Entering edit mode

I may have access to a 16 gig workstation- would that be enough?

0
Entering edit mode

That should be enough

0
Entering edit mode

So you would suggest I use the automatic network construction on a stronger computer versus using the blockwiseModules? If I do that would I need to increase the maximum block size in order to plot more than one module?

0
Entering edit mode

If you have enough RAM for 18k probes you can follow the step-by-step network construction.

Peter Langfelder explains very well why WGCNA analysis on a single block (all genes) is preferred over the blockwiseModules approach: link

0
Entering edit mode

With access to 16gigs of RAM could I use the automatic network construction and increase the maxblocksize to 18k instead of 5000 default?

0
Entering edit mode

you can set maxBlockSize to the total number of genes you have in datExpr