search

farrellja / URD

URD - Reconstruction of Branching Developmental Trajectories
GNU General Public License v3.0
117 stars 41 forks source link

Permanent error by graphClustering #13

Closed fereshtehizadi closed 6 years ago

fereshtehizadi commented 6 years ago

Hi, I have 1562 cells in 8 time point (2h to 16h).

I am running URD, my root stage is h2 and final stage is h16 but I am being stopped at this stage. Could you please help me?

> axial.6somite <- graphClustering(axial.6somite, num.nn = 50, do.jaccard=TRUE, method="Louvain")
Error in graphClustering(axial.6somite, num.nn = 50, do.jaccard = TRUE,  : 
  unused arguments (num.nn = 50, do.jaccard = TRUE, method = "Louvain")
> 

> print(object.6s)

 URD object: 12451 genes x 1556 cells. 

object.6s.mnn <- graphClustering(object.6s) 

Show Traceback Rerun with Debug Error in (function (classes, fdef, mtable) : unable to find an inherited method for function ‘graphClustering’ for signature ‘"URD"’ >

This is whole of my process

> axial <- createURD(count.data = as.matrix(T), meta = r, min.cells=3, min.counts=3)
2018-08-24 09:05:16: Filtering cells by number of genes.
2018-08-24 09:05:18: Filtering genes by number of cells.
2018-08-24 09:05:19: Filtering genes by number of counts across entire data.
2018-08-24 09:05:20: Filtering genes by maximum observed expression.
2018-08-24 09:05:22: Creating URD object.
2018-08-24 09:05:23: Determining normalization factors.
2018-08-24 09:05:24: Normalizing and log-transforming the data.
2018-08-24 09:05:29: Finishing setup of the URD object.
2018-08-24 09:05:29: All done.
> axial@group.ids$stage <- as.character(axial@meta[rownames(axial@group.ids),"stage.nice"])
> stages <- sort(unique(axial@group.ids$stage))

         init stage
s1.1     s1.1   h16
s1.2     s1.2   h16
s1.3     s1.3   h16

> var.by.stage <- lapply(seq(3,8,3), function(n) {
+   findVariableGenes(axial, cells.fit=cellsInCluster(axial, "stage", stages[(n-2):n]), set.object.var.genes=F, diffCV.cutoff=0.3, mean.min=.005, mean.max=100, main.use=paste0("Stages ", stages[n-2], " to ", stages[n]), do.plot=TRUE)
+ })
Warning messages:
1: In xy.coords(x, y, xlabel, ylabel, log) :
  38 x values <= 0 omitted from logarithmic plot
2: In densfun(x, parm[1], parm[2], ...) : NaNs produced
3: In densfun(x, parm[1], parm[2], ...) : NaNs produced
4: In densfun(x, parm[1], parm[2], ...) : NaNs produced
5: In densfun(x, parm[1], parm[2], ...) : NaNs produced
6: In xy.coords(x, y, xlabel, ylabel, log) :
  6 x values <= 0 omitted from logarithmic plot
> var.genes <- sort(unique(unlist(var.by.stage)))
> axial@var.genes <- var.genes
> axial <- calcPCA(axial, mp.factor = 2)
[1] "2018-08-24 09:08:45: Centering and scaling data."
[1] "2018-08-24 09:08:51: Removing genes with no variation."
[1] "2018-08-24 09:08:53: Calculating PCA."
[1] "2018-08-24 09:10:34: Estimating significant PCs."
[1] "Marchenko-Pastur eigenvalue null upper bound: 14.3352330233523"
[1] "9 PCs have eigenvalues larger than 2 times null upper bound."
[1] "Storing 18 PCs."
> pcSDPlot(axial)
> set.seed(19)
> axial <- calcTsne(object = axial)
> plotDim(axial, "stage.nice", plot.title = "tSNE: Stage")
> axial <- calcDM(axial, knn = 40, sigma=16)
[1] "Using provided global sigma 16"
Error in stopifsmall(max(trans_p@x, na.rm = TRUE)) : 
  The supplied sigma is not large enough. Please select a larger one.
            find_sigmas(data) should return one with the right order of magnitude. (max dist. is 5.271e-31)
> axial <- calcDM(axial, knn = 40, sigma=NULL)
[1] "destiny determined an optimal global sigma of 273.752"
> plotDimArray(axial, reduction.use = "dm", dims.to.plot = 1:8, outer.title = "Diffusion Map (Sigma NULL, 40 NNs): Stage", label="stage", plot.title="", legend=F)
> plotDim(axial, "stage.nice", transitions.plot = 10000, plot.title="Developmental stage (with transitions)")
> root.cells <- cellsInCluster(axial, "stage", "h2")
> root.cells <- cellsInCluster(axial, "stage", "h2")
> axial.floods <- floodPseudotime(axial, root.cells = root.cells, n=50, minimum.cells.flooded = 2, verbose=F)
> axial <- floodPseudotimeProcess(axial, axial.floods, floods.name="pseudotime")
> pseudotimePlotStabilityOverall(axial)
> plotDim(axial, "pseudotime")
> plotDists(axial, "pseudotime", "stage", plot.title="Pseudotime by stage")
> axial.6somite <- urdSubset(axial, cells.keep=cellsInCluster(axial, "stage", "h16"))
> axial.6somite@var.genes <- var.by.stage[[4]]
Error in var.by.stage[[4]] : subscript out of bounds
> axial.6somite@var.genes <- var.by.stage[[8]]
Error in var.by.stage[[8]] : subscript out of bounds
> axial.6somite@var.genes <- var.by.stage[[1]]
> axial.6somite <- calcPCA(axial.6somite, mp.factor = 1.5)
[1] "2018-08-24 09:20:30: Centering and scaling data."
[1] "2018-08-24 09:20:31: Removing genes with no variation."
[1] "2018-08-24 09:20:32: Calculating PCA."
[1] "2018-08-24 09:20:34: Estimating significant PCs."
[1] "Marchenko-Pastur eigenvalue null upper bound: 69.0546023898962"
[1] "3 PCs have eigenvalues larger than 1.5 times null upper bound."
[1] "Storing 6 PCs."
> pcSDPlot(axial.6somite)
> set.seed(20)
> axial.6somite <- calcTsne(axial.6somite)
> axial.6somite <- graphClustering(axial.6somite, num.nn = 50, do.jaccard=T, method="Louvain")
Error in graphClustering(axial.6somite, num.nn = 50, do.jaccard = T, method = "Louvain") : 
  unused arguments (num.nn = 50, do.jaccard = T, method = "Louvain")
> axial.6somite <- graphClustering(axial.6somite, num.nn = 50, do.jaccard=TRUE, method="Louvain")
Error in graphClustering(axial.6somite, num.nn = 50, do.jaccard = TRUE,  : 
  unused arguments (num.nn = 50, do.jaccard = TRUE, method = "Louvain")
> axial.6somite <- graphClustering(axial.6somite, do.jaccard=TRUE, method="Louvain")
Error in graphClustering(axial.6somite, do.jaccard = TRUE, method = "Louvain") : 
  unused arguments (do.jaccard = TRUE, method = "Louvain")
> axial.6somite <- graphClustering(axial.6somite, method="Louvain")
Error in graphClustering(axial.6somite, method = "Louvain") : 
  unused argument (method = "Louvain")
> axial.6somite <- graphClustering(axial.6somite)
 Show Traceback

 Rerun with Debug
 Error in (function (classes, fdef, mtable)  : 
  unable to find an inherited method for function ‘graphClustering’ for signature ‘"URD"’ > plotDim(axial.6somite, "Louvain-50", plot.title = "Louvain (50 NN) graph clustering", point.size=3)
Error in data.for.plot(object, label = label, label.type = label.type,  : 
  Cannot find Louvain-50 in metadata, group.ids, signatures, genes, or pseudotime.
> View(read.table("meta.axial.txt"))
> axial.6somite@var.genes <- var.by.stage[[2]]
> axial.6somite <- calcPCA(axial.6somite, mp.factor = 1.5)
[1] "2018-08-24 09:24:13: Centering and scaling data."
[1] "2018-08-24 09:24:15: Removing genes with no variation."
[1] "2018-08-24 09:24:15: Calculating PCA."
[1] "2018-08-24 09:24:17: Estimating significant PCs."
[1] "Marchenko-Pastur eigenvalue null upper bound: 71.5558152464193"
[1] "3 PCs have eigenvalues larger than 1.5 times null upper bound."
[1] "Storing 6 PCs."
> pcSDPlot(axial.6somite)
> set.seed(20)
> axial.6somite <- calcTsne(axial.6somite)
> axial.6somite@var.genes <- var.by.stage[[2]]
> axial.6somite <- graphClustering(axial.6somite, num.nn = 50, do.jaccard=TRUE, method="Louvain")
Error in graphClustering(axial.6somite, num.nn = 50, do.jaccard = TRUE,  : 
  unused arguments (num.nn = 50, do.jaccard = TRUE, method = "Louvain")
> axial.6somite@var.genes <- var.by.stage[[3]]
Error in var.by.stage[[3]] : subscript out of bounds
> axial.6somite@var.genes <- var.by.stage[[4]]
Error in var.by.stage[[4]] : subscript out of bounds
> axial.6somite@var.genes <- var.by.stage[[5]]
Error in var.by.stage[[5]] : subscript out of bounds
> axial.6somite@var.genes <- var.by.stage[[6]]
Error in var.by.stage[[6]] : subscript out of bounds
> 7# Calculate PCA and tSNE
[1] 7
> axial.6somite@var.genes <- var.by.stage[[7]]
Error in var.by.stage[[7]] : subscript out of bounds
> axial.6somite@var.genes <- var.by.stage[[8]]
Error in var.by.stage[[8]] : subscript out of bounds
> axial.6somite@var.genes <- var.by.stage[[1]]
> axial.6somite <- calcPCA(axial.6somite, mp.factor = 1.5)
[1] "2018-08-24 09:26:14: Centering and scaling data."
[1] "2018-08-24 09:26:15: Removing genes with no variation."
[1] "2018-08-24 09:26:16: Calculating PCA."
[1] "2018-08-24 09:26:18: Estimating significant PCs."
[1] "Marchenko-Pastur eigenvalue null upper bound: 69.0546023898962"
[1] "3 PCs have eigenvalues larger than 1.5 times null upper bound."
[1] "Storing 6 PCs."
> pcSDPlot(axial.6somite)
> set.seed(20)
> axial.6somite <- calcTsne(axial.6somite)
> axial.6somite <- graphClustering(axial.6somite, num.nn = 50, do.jaccard=TRUE, method="Louvain")
Error in graphClustering(axial.6somite, num.nn = 50, do.jaccard = TRUE,  : 
  unused arguments (num.nn = 50, do.jaccard = TRUE, method = "Louvain")
>
fereshtehizadi commented 6 years ago

URD::graphClustering solved that