Seurat: ReadH5AD error: object of type 'environment' is not subsettable
When I used ReadH5AD to read a .h5ad file generated from scanpy, there was an error:
Pulling expression matrices and metadata
Data is unscaled
Error in file[["obs"]][]: object of type 'environment' is not subsettable
Traceback:
and my code is adata = ReadH5AD(file ="../adata_raw.h5ad",verbose = TRUE)
adata_raw.h5ad information:
AnnData object with n_obs × n_vars = 52049 × 19760
obs: 'donor', 'disease', 'tissue', 'celltype', 'name', 'protocol', 'gse_id', 'n_genes'
var: 'n_cells'
When I used ReadH5AD to read a smaller adata.h5ad(15704 x 14895) file, it did not report an error.
AnnData object with n_obs × n_vars = 15704 × 14895
obs: 'batch', 'celltype', 'disease', 'donor', 'library', 'name', 'protocol', 'gse_id', 'n_genes'
var: 'n_cells'
Does anyone give me some advice? Thanks!
tkisss
All 37 comments
I'm having the same error:
Error in file[["obs"]][]: object of type 'environment' is not subsettable
sessionInfo()
R version 3.5.1 (2018-07-02)
Platform: x86_64-pc-linux-gnu (64-bit)
Running under: CentOS Linux 7 (Core)
Matrix products: default
BLAS/LAPACK: /n/app/openblas/0.2.19/lib/libopenblas_core2p-r0.2.19.so
locale:
[1] LC_CTYPE=en_US.UTF-8 LC_NUMERIC=C
[3] LC_TIME=en_US.UTF-8 LC_COLLATE=en_US.UTF-8
[5] LC_MONETARY=en_US.UTF-8 LC_MESSAGES=en_US.UTF-8
[7] LC_PAPER=en_US.UTF-8 LC_NAME=C
[9] LC_ADDRESS=C LC_TELEPHONE=C
[11] LC_MEASUREMENT=en_US.UTF-8 LC_IDENTIFICATION=C
attached base packages:
[1] stats graphics grDevices utils datasets methods base
other attached packages:
[1] Seurat_3.0.2
loaded via a namespace (and not attached):
[1] nlme_3.1-137 tsne_0.1-3 bitops_1.0-6
[4] bit64_0.9-7 RColorBrewer_1.1-2 httr_1.4.1
[7] sctransform_0.2.1 tools_3.5.1 backports_1.1.5
[10] R6_2.4.0 irlba_2.3.3 KernSmooth_2.23-15
[13] lazyeval_0.2.2 colorspace_1.4-1 npsurv_0.4-0
[16] tidyselect_0.2.5 gridExtra_2.3 bit_1.1-14
[19] compiler_3.5.1 hdf5r_1.3.0 plotly_4.9.1
[22] caTools_1.17.1.2 scales_1.0.0 lmtest_0.9-37
[25] ggridges_0.5.1 pbapply_1.4-2 stringr_1.4.0
[28] digest_0.6.22 R.utils_2.9.0 pkgconfig_2.0.3
[31] htmltools_0.4.0 bibtex_0.4.2 htmlwidgets_1.5.1
[34] rlang_0.4.1 zoo_1.8-6 jsonlite_1.6
[37] ica_1.0-2 gtools_3.8.1 dplyr_0.8.3
[40] R.oo_1.23.0 magrittr_1.5 Matrix_1.2-17
[43] Rcpp_1.0.3 munsell_0.5.0 ape_5.3
[46] reticulate_1.13 lifecycle_0.1.0 R.methodsS3_1.7.1
[49] stringi_1.4.3 gbRd_0.4-11 MASS_7.3-51.4
[52] gplots_3.0.1.1 Rtsne_0.15 plyr_1.8.4
[55] grid_3.5.1 parallel_3.5.1 gdata_2.18.0
[58] listenv_0.7.0 ggrepel_0.8.1 crayon_1.3.4
[61] lattice_0.20-38 cowplot_1.0.0 splines_3.5.1
[64] SDMTools_1.1-221.1 zeallot_0.1.0 pillar_1.4.2
[67] igraph_1.2.4.1 future.apply_1.3.0 reshape2_1.4.3
[70] codetools_0.2-16 glue_1.3.1 lsei_1.2-0
[73] metap_1.1 data.table_1.12.6 vctrs_0.2.0
[76] png_0.1-7 Rdpack_0.11-0 gtable_0.3.0
[79] RANN_2.6.1 purrr_0.3.3 tidyr_1.0.0
[82] future_1.15.0 assertthat_0.2.1 ggplot2_3.2.1
[85] rsvd_1.0.2 survival_3.1-6 viridisLite_0.3.0
[88] tibble_2.1.3 cluster_2.0.7-1 globals_0.12.4
[91] fitdistrplus_1.0-14 ROCR_1.0-7
sergiopolid
on 22 Jan 2020
Not sure if this is the fix, but i faced the same error but the error was because the h5ad was constructed in scanpy 1.4.5 . when i downgraded to scanpy 1.4.4.post1 ( which worked fine before) and recreated the h5ad, there is no error
lsundaram
on 27 Jan 2020
Also getting this error, on both scanpy 1.4.5 and scanpy 1.4.4.post1 on Seurat 3.0.2
thomas-davis
on 28 Jan 2020
Hoping to up the priority for this fix. I'm also receiving the error for scanpy 1.4.5 and Seurat 3.0.2.
apblair
on 5 Feb 2020
I was having this problem until few minutes ago.
Some files were working with ReadH5AD, some weren't. I figured out that the files which were working were saved with an older version of of anndata.
So I downgraded anndata-0.7.1 to 0.6.22.post1, recomputed the data and overwrote the files. Now seurat reads it properly. I'm using Seurat-3.1.2 btw.
lrafa
on 5 Feb 2020
I'm also getting this error for scanpy 1.4.5 and Seurat 3.1.0:
Pulling expression matrices and metadata
Data is scaled
Error in file[["obs"]][] :
object of type 'environment' is not subsettable
I tried changing my install of anndata by running:
pip install anndata==0.6.22.post1
...and then re-exporting the file and re-testing ReadH5AD.
I then get this new output:
Pulling expression matrices and metadata
Data is scaled
Error in UseMethod(generic = "as.sparse", object = x) :
no applicable method for 'as.sparse' applied to an object of class "c('H5D', 'H5RefClass', 'R6')"
Installed X11, upgraded to Seurat 3.1.2 using
devtools::install_version(package='Seurat', version=package_version('3.1.2')
But I'm still getting the same 'as.sparse' error from above when I try to use H5AD.
Hope this helps
kmuench
on 11 Feb 2020
@kmuench
Did you copy the anndata data (while it was still sparse) into the .raw attribute before scaling and saving it? E.g.
adata = load_sparse_data
adata.raw = adata
sc.pp.scale(adata)
adata.write('myfile.h5ad')
lrafa
on 11 Feb 2020
Ahh, I didn't explicitly copy anything into raw, but I'm wondering if my issue is that I'm importing one of the test datasets? (Development of Myeloid Progenitors)
I'm trying to save the output of the DCA pipeline (https://github.com/theislab/dca) which may not be affecting the same data slots as pp.scale:
import scanpy as sc
paul15_ae = sc.datasets.paul15()
sc.pp.filter_genes(paul15_ae, min_counts=1)
dca(paul15_ae, threads=1)
sc.pp.normalize_per_cell(paul15_ae)
sc.pp.log1p(paul15_ae)
sc.pp.pca(paul15_ae)
import anndata as an
an.AnnData.write(paul15_ae, filename = 'paul15_ae.h5ad')
Hi, has anyone come up with a solution to this? I'm having the same issue.
asn32
on 15 Feb 2020
same issue.
tiancaigg
on 25 Feb 2020
I found if you set ReadH5AD's argument layers=FALSE the error goes away.
apblair
on 26 Feb 2020
Unfortunately not for me.
I still get
adata = ReadH5AD(file ="./adata.h5ad", layers=FALSE)
Pulling expression matrices and metadata
Data is unscaled
Error in file[["obs"]][] :
object of type 'environment' is not subsettable
Acribbs
on 26 Feb 2020
Apologies for the delay in responding. Please can you post the h5ad file (you can post a link for us to download, we would suggest making the object as small as possible to facilitate debugging), and we will take a look
satijalab
on 5 Mar 2020
@satijalab
This toy dataset recreates the error -->
(https://www.dropbox.com/s/c6kz82kcy6w0wbw/test.h5ad?dl=0)
Thanks in advance!
Errror ->
Error in file[["obs"]][]: object of type 'environment' is not subsettable
Traceback:
1. ReadH5AD(file = "../../data/test.h5ad")
2. ReadH5AD.character(file = "../../data/test.h5ad")
3. ReadH5AD(file = hfile, assay = assay, layers = layers, verbose = verbose,
. ...)
4. ReadH5AD.H5File(file = hfile, assay = assay, layers = layers,
. verbose = verbose, ...)
franciscogrisanti
on 6 Mar 2020
Regarding #2373,
This line seems to raise the error: https://github.com/satijalab/seurat/blob/94343c4fdb35a3a1e7a7e14d4a0bb0cec657075c/R/objects.R#L2657
There is a mismatch of dimension shapes between indptr and rev(x = hdf5r::h5attr(x = x, which = 'h5sparse_shape'))
> file[['raw.X']]
Class: H5Group
Group: /raw.X
Attributes: h5sparse_format, h5sparse_shape
Listing:
name obj_type dataset.dims dataset.type_class
data H5I_DATASET 33518842 H5T_FLOAT
indices H5I_DATASET 33518842 H5T_INTEGER
indptr H5I_DATASET 8960 H5T_INTEGER
> rev(x = hdf5r::h5attr(x = file[['raw.X']], which = 'h5sparse_shape'))
[1] 8959 12241
md5-4cedd8636d2e2a094f5955bab93acf03
sparseMat <- Matrix::sparseMatrix(
i = x[['indices']][] + 1,
p = x[['indptr']][],
x = x[['data']][]
)
Although,I currently don't have a clue why the dimension shapes do not match
N.B.: Loading the same H5AD file in Scanpy works fine
dweemx
on 6 Mar 2020
Hi @dweemx
I tried that fix but it didn't work out either. Could you check my approach to see if it match yours?
Thanks in advance!
I created a custom "as.sparse.H5Group"
custom_as.sparse.H5Group <- function(x, ...) {
CheckDots(...)
for (i in c('data', 'indices', 'indptr')) {
if (!x$exists(name = i) || !is(object = x[[i]], class2 = 'H5D')) {
stop("Invalid H5Group specification for a sparse matrix, missing dataset ", i)
}
}
if ('h5sparse_shape' %in% hdf5r::h5attr_names(x = x)) {
return(Matrix::sparseMatrix(
i = x[['indices']][] + 1,
p = x[['indptr']][],
x = x[['data']][]
#dims = rev(x = hdf5r::h5attr(x = file[['raw.X']], which = 'h5sparse_shape'))
))
}
return(Matrix::sparseMatrix(
i = x[['indices']][] + 1,
p = x[['indptr']][],
x = x[['data']][]
))
}
Set the function up and running
environment(custom_as.sparse.H5Group) <- asNamespace('Seurat')
assignInNamespace("as.sparse.H5Group", custom_as.sparse.H5Group, ns = "Seurat")
Check is correct
getAnywhere(as.sparse.H5Group)
A single object matching ‘as.sparse.H5Group’ was found
It was found in the following places
registered S3 method for as.sparse from namespace Seurat
namespace:Seurat
with value
function(x, ...) {
CheckDots(...)
for (i in c('data', 'indices', 'indptr')) {
if (!x$exists(name = i) || !is(object = x[[i]], class2 = 'H5D')) {
stop("Invalid H5Group specification for a sparse matrix, missing dataset ", i)
}
}
if ('h5sparse_shape' %in% hdf5r::h5attr_names(x = x)) {
return(Matrix::sparseMatrix(
i = x[['indices']][] + 1,
p = x[['indptr']][],
x = x[['data']][]
#dims = rev(x = hdf5r::h5attr(x = file[['raw.X']], which = 'h5sparse_shape'))
))
}
return(Matrix::sparseMatrix(
i = x[['indices']][] + 1,
p = x[['indptr']][],
x = x[['data']][]
))
}
<environment: namespace:Seurat>
But I get again while loading -->
test <- ReadH5AD(file = '../../data/test.h5ad')
Error in file[["obs"]][]: object of type 'environment' is not subsettable
Traceback:
1. ReadH5AD(file = "../../data/test.h5ad")
2. ReadH5AD.character(file = "../../data/test.h5ad")
3. ReadH5AD(file = hfile, assay = assay, layers = layers, verbose = verbose,
. ...)
4. ReadH5AD.H5File(file = hfile, assay = assay, layers = layers,
. verbose = verbose, ...)
Hey @franciscogrisanti,
I just tested your solution, it's working for me!
dweemx
on 9 Mar 2020
Thanks for the prompt response @dweemx
Could you provide me your scanpy/seurat version?
Thanks again!
Mine are ->
For scanpy
scanpy==1.4.5.1 anndata==0.7.1 umap==0.3.10 numpy==1.17.3 scipy==1.4.1 pandas==0.25.3 scikit-learn==0.22.1 statsmodels==0.11.0 python-igraph==0.7.1 louvain==0.6.1
For Seurat
R version 3.6.1 (2019-07-05)
Platform: x86_64-apple-darwin13.4.0 (64-bit)
Running under: macOS Catalina 10.15
Matrix products: default
BLAS/LAPACK: /Users/franciscogrisanti/anaconda3/envs/seurat/lib/R/lib/libRblas.dylib
locale:
[1] en_US.UTF-8/en_US.UTF-8/en_US.UTF-8/C/en_US.UTF-8/en_US.UTF-8
attached base packages:
[1] parallel stats4 stats graphics grDevices utils datasets
[8] methods base
other attached packages:
[1] patchwork_1.0.0 loomR_0.2.1.9000
[3] hdf5r_1.3.1 R6_2.4.1
[5] scater_1.14.0 ggplot2_3.2.1
[7] SingleCellExperiment_1.8.0 SummarizedExperiment_1.16.0
[9] DelayedArray_0.12.0 BiocParallel_1.20.0
[11] matrixStats_0.55.0 Biobase_2.46.0
[13] GenomicRanges_1.38.0 GenomeInfoDb_1.22.0
[15] IRanges_2.20.0 S4Vectors_0.24.0
[17] BiocGenerics_0.32.0 Seurat_3.1.4
loaded via a namespace (and not attached):
[1] Rtsne_0.15 ggbeeswarm_0.6.0 colorspace_1.4-1
[4] ggridges_0.5.2 IRdisplay_0.7.0 XVector_0.26.0
[7] base64enc_0.1-3 BiocNeighbors_1.4.0 leiden_0.3.3
[10] listenv_0.8.0 npsurv_0.4-0 bit64_0.9-7
[13] ggrepel_0.8.1 codetools_0.2-16 splines_3.6.1
[16] lsei_1.2-0 IRkernel_1.1 jsonlite_1.6.1
[19] ica_1.0-2 cluster_2.1.0 png_0.1-7
[22] uwot_0.1.5 sctransform_0.2.1 compiler_3.6.1
[25] httr_1.4.1 assertthat_0.2.1 Matrix_1.2-18
[28] lazyeval_0.2.2 BiocSingular_1.2.0 htmltools_0.4.0
[31] tools_3.6.1 rsvd_1.0.2 igraph_1.2.4.2
[34] gtable_0.3.0 glue_1.3.1 GenomeInfoDbData_1.2.2
[37] RANN_2.6.1 reshape2_1.4.3 dplyr_0.8.0.1
[40] Rcpp_1.0.3 gdata_2.18.0 ape_5.3
[43] nlme_3.1-145 DelayedMatrixStats_1.8.0 gbRd_0.4-11
[46] lmtest_0.9-37 stringr_1.4.0 globals_0.12.5
[49] irlba_2.3.3 gtools_3.8.1 future_1.16.0
[52] MASS_7.3-51.5 zlibbioc_1.32.0 zoo_1.8-7
[55] scales_1.0.0 RColorBrewer_1.1-2 reticulate_1.14
[58] pbapply_1.4-2 gridExtra_2.3 stringi_1.4.3
[61] caTools_1.18.0 bibtex_0.4.2.2 repr_1.1.0
[64] Rdpack_0.11-1 rlang_0.4.5 pkgconfig_2.0.3
[67] bitops_1.0-6 evaluate_0.14 lattice_0.20-40
[70] ROCR_1.0-7 purrr_0.3.3 htmlwidgets_1.5.1
[73] bit_1.1-14 cowplot_1.0.0 tidyselect_0.2.5
[76] RcppAnnoy_0.0.12 plyr_1.8.6 magrittr_1.5
[79] gplots_3.0.3 pbdZMQ_0.3-3 pillar_1.4.3
[82] withr_2.1.2 fitdistrplus_1.0-14 survival_3.1-8
[85] RCurl_1.95-4.12 tibble_2.1.3 future.apply_1.4.0
[88] tsne_0.1-3 crayon_1.3.4 uuid_0.1-2
[91] KernSmooth_2.23-15 plotly_4.9.2 viridis_0.5.1
[94] grid_3.6.1 data.table_1.12.2 metap_1.1
[97] digest_0.6.25 tidyr_0.8.3 RcppParallel_4.4.4
[100] munsell_0.5.0 beeswarm_0.2.3 viridisLite_0.3.0
[103] vipor_0.4.5
Mine are these:
Scanpy
anndata==0.6.22.post1
annoy==1.15.2
numpy==1.17.2
pandas==0.25.1
scanpy==1.4.4.post1
scikit-learn==0.21.3
scipy==1.3.1
sklearn==0.0
statsmodels==0.10.1
umap-learn==0.3.10
Seurat
R version 3.6.0 (2019-04-26)
Platform: x86_64-pc-linux-gnu (64-bit)
Running under: CentOS Linux 7 (Core)
Matrix products: default
BLAS/LAPACK:~/software/OpenBLAS/0.2.20-GCC-6.4.0-2.28/lib/libopenblas_haswellp-r0.2.20.so
Random number generation:
RNG: Mersenne-Twister
Normal: Inversion
Sample: Rounding
locale:
[1] LC_CTYPE=en_US.UTF-8 LC_NUMERIC=C LC_TIME=en_US.UTF-8 LC_COLLATE=en_US.UTF-8 LC_MONETARY=en_US.UTF-8
[6] LC_MESSAGES=en_US.UTF-8 LC_PAPER=en_US.UTF-8 LC_NAME=C LC_ADDRESS=C LC_TELEPHONE=C
[11] LC_MEASUREMENT=en_US.UTF-8 LC_IDENTIFICATION=C
attached base packages:
[1] parallel stats4 stats graphics grDevices utils datasets methods base
other attached packages:
[1] IRanges_2.18.2 S4Vectors_0.22.0 Biobase_2.44.0 BiocGenerics_0.30.0
[7] Seurat_3.1.2 reticulate_1.14
loaded via a namespace (and not attached):
[1] nlme_3.1-139 tsne_0.1-3 bitops_1.0-6 bit64_0.9-7 RcppAnnoy_0.0.12 RColorBrewer_1.1-2
[7] httr_1.4.1 sctransform_0.2.0 tools_3.6.0 backports_1.1.4 R6_2.4.0 irlba_2.3.3
[13] KernSmooth_2.23-15 DBI_1.0.0 uwot_0.1.5 lazyeval_0.2.2 colorspace_1.4-1 npsurv_0.4-0
[19] tidyselect_0.2.5 gridExtra_2.3 bit_1.1-14 compiler_3.6.0 hdf5r_1.2.0 plotly_4.9.0
[25] caTools_1.17.1.2 scales_1.0.0 lmtest_0.9-37 ggridges_0.5.1 pbapply_1.4-0 rappdirs_0.3.1
[31] stringr_1.4.0 digest_0.6.19 R.utils_2.9.0 pkgconfig_2.0.2 htmltools_0.3.6 bibtex_0.4.2
[37] htmlwidgets_1.3 rlang_0.4.0 RSQLite_2.1.2 rstudioapi_0.10 zoo_1.8-6 jsonlite_1.6
[43] ica_1.0-2 gtools_3.8.1 dplyr_0.8.2 R.oo_1.22.0 magrittr_1.5 Matrix_1.2-17
[49] Rcpp_1.0.1 munsell_0.5.0 ape_5.3 lifecycle_0.1.0 R.methodsS3_1.7.1 stringi_1.4.3
[55] gbRd_0.4-11 MASS_7.3-51.4 gplots_3.0.1.1 Rtsne_0.15 plyr_1.8.4 blob_1.2.0
[61] grid_3.6.0 gdata_2.18.0 listenv_0.7.0 ggrepel_0.8.1 crayon_1.3.4 lattice_0.20-38
[67] cowplot_0.9.4 splines_3.6.0 SDMTools_1.1-221.1 zeallot_0.1.0 pillar_1.4.2 igraph_1.2.4.2
[73] future.apply_1.3.0 reshape2_1.4.3 codetools_0.2-16 leiden_0.3.1 glue_1.3.1 lsei_1.2-0
[79] metap_1.1 RcppParallel_4.4.3 data.table_1.12.2 vctrs_0.2.0 png_0.1-7 Rdpack_0.11-0
[85] gtable_0.3.0 RANN_2.6.1 purrr_0.3.2 tidyr_1.0.0 future_1.14.0 assertthat_0.2.1
[91] ggplot2_3.2.1 rsvd_1.0.2 survival_2.44-1.1 viridisLite_0.3.0 tibble_2.1.3 memoise_1.1.0
[97] cluster_2.0.8 globals_0.12.4 fitdistrplus_1.0-14 ROCR_1.0-7
Seems that you are using the latest version of scanpy and anndata
dweemx
on 9 Mar 2020
Thank you so much @dweemx I was able to load the H5AD object into seurat 😁
franciscogrisanti
on 9 Mar 2020
So what was the issue ? Did you fix it by downgrading scanpy and anndata ?
dweemx
on 9 Mar 2020
I think the issue is in h5py==2.10.0. Any .h5ad object exported with this version breaks while loading into Seurat 3.4.1. I then exported my objects as .loom
Downgraded to h5py==2.9.0, loaded the .loom, exported as .h5ad again. And finally, I was able to read them in Seurat.
I had to did the .loom trick, because apparently there is no backward compatibility between h5py==2.10.0 and h5py==2.9.0 for the h5ad files either.
franciscogrisanti
on 9 Mar 2020
My issue was resolved with downgrading h5py to 2.9.0 too and remaking the .h5ad files. Thanks for the solution.
Acribbs
on 9 Mar 2020
Downgrading h5py to 2.9.0 and remaking the .5had file did not fix the issue for me.
@satijalab any progress regarding this bug?
kleurless
on 18 Mar 2020
I think the issue is in h5py==2.10.0. Any .h5ad object exported with this version breaks while loading into Seurat 3.4.1. I then exported my objects as .loom
Downgraded to h5py==2.9.0, loaded the .loom, exported as .h5ad again. And finally, I was able to read them in Seurat.
I had to did the .loom trick, because apparently there is no backward compatibility between h5py==2.10.0 and h5py==2.9.0 for the h5ad files either.
Neither did it work for me. Did you downgrade scanpy or anndata?
hejian41
on 19 Mar 2020
This issue has been automatically closed because there has been no response to our request for more information from the original author. With only the information that is currently in the issue, we don't have enough information to take action. Please reach out if you have or find the answers we need so that we can investigate further.
![no-response[bot] picture](https://avatars2.githubusercontent.com/u/26350515?v=4&s=40)
no-response[bot]
on 19 Mar 2020
I installed all my software using Conda. Scanpy==1.4.4.post1, anndata==0.6.22.post1.
Acribbs
on 20 Mar 2020
I decided to export to scipy.io.mmwrite and create the three files similar to 10x.
Puriney
on 30 Mar 2020
Currently it seems that the latest version of Scanpy doesn't produce a h5ad file that Seurat can read. The layers=FALSE trick also doesn't work for me. Is the Seurat team going to change the h5ad reading method or we have to use older versions of Scanpy?
brianpenghe
on 23 May 2020
This issue is a major blocker
sjfleming
on 10 Jun 2020
Using the library SeuratDisk worked for me with the latest versions of Seurat and Scanpy.
https://github.com/satijalab/seurat/issues/2748#issuecomment-629461886
Manikgarg
on 20 Jun 2020
I'm having the same issue. Setting layers=FALSE did not work for me. If I downgrade to h5py==2.9.0, scanpy==1.4.4.post1 and anndata==0.6.22.post1 , I get a different error:
```
Error in validObject(r) :
invalid class “dgTMatrix” object: all row indices (slot 'i') must be between 0 and nrow-1 in a TsparseMatrix
astair
on 4 Aug 2020
using SeuratDisk worked for me :)
SBata
on 16 Aug 2020
SeuratDisk unfortunately didn't work out for me.
I posted my problem in the following issue.
https://github.com/mojaveazure/seurat-disk/issues/18
does anyone face a similar issue?
Averiian
on 17 Aug 2020
SeuratDisk unfortunately didn't work out for me.
I posted my problem in the following issue.does anyone face a similar issue?
SeuratDisk did not work for me either.
The .hda5 file that I am using is here. It was generated in this blog post where the blog writer has video links to how they converted multiple .tar.gz and .csv files into this one file.
I am not familiar enough with scanpy to be able to recreate this file quickly.
ColetheStatistician
on 4 Sep 2020
Hi,
I found a work around this using reticulate, mostly modified from here . Though its not straight forward, but this works for me!. Keeping it as a reference for others!.
suppressPackageStartupMessages({
library("reticulate")
library("ggplot2")
library("scater")
library("Seurat")
})
Import scanpy and object
sc <- import("scanpy")
adata <- sc$read_h5ad("myAnndata.h5ad")
adata
````
AnnData object with n_obs × n_vars = 21391 × 30983
obs: 'n_genes', 'n_counts', 'percent_mito', 'leiden_res_1.3', 'leiden_res_2', 'leiden_labels', 'leiden_res_0.9', 'leiden_res_1.1', 'leiden_res_0.5'
var: 'n_cells', 'percent_cells', 'robust', 'highly_variable_features', 'mean', 'var', 'hvf_loess', 'hvf_rank'
uns: 'Channels', 'PCs', 'fmat_highly_variable_features', 'leiden_labels_colors', 'leiden_res_0.5_colors', 'leiden_res_0.9_colors', 'leiden_res_1.1_colors', 'pca'
obsm: 'X_pca', 'X_umap'
varm: 'PCs', 'de_res'
layers: 'counts', 'winsorized'
obsp: 'connectivities', 'distances'
## Format count and normalized data for their respective slots
counts <- t(adata$layers["counts"])
colnames(counts) <- adata$obs_names$to_list()
rownames(counts) <- adata$var_names$to_list()
counts <- Matrix::Matrix(as.matrix(counts), sparse = T)
data <- t(adata$layers["winsorized"])
colnames(data) <- adata$obs_names$to_list()
rownames(data) <- adata$var_names$to_list()
data <- Matrix::Matrix(as.matrix(data), sparse = T)
## Create the Seurat object
seurat <- CreateSeuratObject(counts)
seurat <- SetAssayData(seurat, "data", data)
seurat <- AddMetaData(seurat, adata$obs)
## Import dim. reductions
embedding <- adata$obsm["X_umap"]
rownames(embedding) <- adata$obs_names$to_list()
colnames(embedding) <- c("umap_1", "umap_2")
seurat[["umap"]] <- CreateDimReducObject(embedding, key = "umap_")
seurat
An object of class Seurat
30983 features across 21391 samples within 1 assay
Active assay: RNA (30983 features, 0 variable features)
1 dimensional reduction calculated: umap
kvshams
on 4 Sep 2020
Hi @kvshams, thanks for the detailed example. I am trying to follow your code and getting an error at the "Import scanpy and object" step.
Error in py_get_attr_impl(x, name, silent) :
AttributeError: module 'scanpy' has no attribute 'read_h5ad'
Any thoughts on how to get past this error ? Thanks.
aditisk
on 11 Nov 2020
Related issues
farhanma
·
3Comments
sarahwajid
·
3Comments
htc502
·
3Comments
akhst7
·
3Comments
GHAStVHenry
·
3Comments
Most helpful comment
This issue is a major blocker