> Dear @yuGithuuub,

[Emma-China]

Dear @yuGithuuub,

I currently do not have plans to integrate Harmony into the full Monocle framework. In my own analyses, I found that the trajectory inference algorithm used by Monocle, DDRTree, works just as well on a cell-by-gene matrix as a cell-by-PCs matrix. So I use DDRTree directly on my Harmonized PCA matrix to return the trajectory embedding.

library(DDRTree)
get_ddr <- function(harmony_Z, ndim=2) {
    ncells <- nrow(harmony_Z)
    ncenter <- round(2 * 100 * log(ncells) / (log(ncells) + log(100)))
    ddr_args <- c(list(
        X = harmony_Z, 
        dimensions = ndim, ## LOW DIMENSIONALITY
        ncenter = ncenter, 
        verbose = FALSE))
    ddrtree_res <- do.call(DDRTree, ddr_args)
    ddrtree_res$Z <- t(ddrtree_res$Z)
    row.names(ddrtree_res$Z) <- row.names(harmony_Z)
    colnames(ddrtree_res$Z) <- colnames(harmony_Z)
    return(ddrtree_res$Z)
}

In the code above, harmony_Z is a cell-by-PC matrix of Harmonized cell embeddings and ndim is the number of dimensions you want to get from the DDRTree inference. This is not a full trajectory analysis but it can help you to visualize your cells in a "trajectory space."

Best, Ilya

Hello,I have encountered a problem when i run the function of get_ddr,and would like to ask. Here are my code and error:

harmony_embedding <- Seurat::Embeddings(seuratObject, 'harmony') harmony_embedding[seq_len(5), seq_len(5)] harmony_1 harmony_2 harmony_3 harmony_4 harmony_5 Q9D5_AAACCCAAGCACTAAA-1 1.819167 -7.420685 -6.262630 0.3244117 0.9872209 Q9D5_AAACCCAAGCGATCGA-1 -9.997331 13.725940 4.829441 -13.7063330 3.9843493 Q9D5_AAACCCACAGTAACAA-1 -4.230632 -9.308852 -3.642947 0.1326170 -1.7054477 Q9D5_AAACGAAAGGATTCAA-1 -8.199931 -10.121518 -3.336982 1.8577525 -1.6456189 Q9D5_AAACGAAAGTCATAGA-1 -5.126308 -10.118114 -4.421629 0.4208384 -1.4270345 get_ddr <- function(harmony_Z, ndim=2) { ncells <- nrow(harmony_Z) ncenter <- round(2 * 100 * log(ncells) / (log(ncells) + log(100))) ddr_args <- c(list( X = harmony_Z, dimensions = ndim, ## LOW DIMENSIONALITY ncenter = ncenter, verbose = FALSE)) ddrtree_res <- do.call(DDRTree, ddr_args) ddrtree_res$Z <- t(ddrtree_res$Z) row.names(ddrtree_res$Z) <- row.names(harmony_Z) colnames(ddrtree_res$Z) <- colnames(harmony_Z) return(ddrtree_res$Z) } get_ddr(harmony_embedding) Error in (function (X, dimensions = 2, initial_method = NULL, maxIter = 20, : Error: ncenters must be greater than or equal to ncol(X)

How can I solve such a problem? Thanks for your answer !

Originally posted by @Emma-1991 in https://github.com/immunogenomics/harmony/issues/74#issuecomment-725076069

[ilyakorsunsky]

Dear @Emma-1991,

From the error message, it sounds like there may be an issue with the dimension of the matrix passed to DDRTree. Have you tried to transpose it first?

Best, Ilya