Troubleshooting Converting Seurat Object to Hummus Object with 10X-Multiome Dataset

[cantinilab]

Troubleshooting Converting Seurat Object to Hummus Object with 10X-Multiome Dataset

[r-trimbour]

Dear Daniel,

Sorry for late answer, I was off then a lot of work to catch up. I hope everything is now fine about your family matters. :)

I finally identified the problem with SeuratV5, thanks a lot for pointing out this incompatibility. I should update HuMMuS in the week and let you know as soon as I'm done so you can test it ! :)

Have a great day,
Rémi

[inhyeoklee]

It works now! What was the issue with Seurat V5 if you don't mind sharing?

Thanks,
Daniel

[r-trimbour]

So in the previous version, Hummus objects were built on top of Seurat objects, which allowed to apply all Seurat functions to it.

If I understood correctly the update, with Seurat V5 they changed the way of how their functions extract the names of the assays stored in the SeuratObject. They are now checking before running the functions if it's a SeuratObject and its version, which was annoying when creating a new type of class on top of it.
--> Error in validObject(object = .Object) : invalid class "Seurat" object: 'assays' must be a named list.

To solve this but keep HuMMuS usable with both Seurat package version, I chose made Hummus_Object class more independent from SeuratObjects now.

I added a Hummus_Object@assays slot, that will contain only the Seurat Assays.
I then added/adapted some methods to work with this @assays slot instead of the whole previously stored Seurat Object

It should give lighter Hummus_Objects and reduce compatibility issues as long as they don't modify too much Assays stuctures. 🙂

[inhyeoklee]

Ah What a clever workaround! Thanks for your work, Rémi. :)

[r-trimbour]

Dear Daniel,

Thanks you for your interest in HuMMuS !

This error might be related to the versions of Seurat and Signac. Is it a Seurat v5 object ?
HuMMuS has been developped mostly using Seurat v4, so it might still have some difficulties with seurat v5 objects.

You can maybe give a first try to start a new RSession, then precise options(Seurat.object.assay.version = 'v4') at the beginning of your code.
You can then reduce the seurat object you're passing to hummus to only the relevant assays for HuMMuS to limit unknown factors in the debugging process :
new_seurat <- SeuratObject::CreateSeuratObject(pbmc[['RNA']]@counts)
new_seurat[['peaks']] <- Signac::CreateChromatinAssay(pbmc[['peaks']]@counts (or 'ATAC' depending of the one relevant for you here)

Let me know if enforcing the Seurat assay version works ! :)
Otherwise you can load both Signac and SeuratObject, and post here the sessionInfo() and SeuratObject::Version(pbmc) outputs so we can explore it a bit more.

Best,
Rémi

[inhyeoklee]

Hi Rémi, I apologize for my delayed response here (had some urgent family matter).

I ran the scripts as you suggested, but I'm still getting the same error when I eventually get to creating a hummus object stage. It turns out when I run "SeuratObject::Version(pbmc)", my Seurat object version is still "5.0.0". I'm thinking about downgrading library versions for both Signac and Seurat and then give another try at downgrading my Seurat object. I'll also share my session info later today.

[inhyeoklee]

Here's my session info:
R version 4.3.2 (2023-10-31)
Platform: x86_64-conda-linux-gnu (64-bit)
Running under: CentOS Linux 7 (Core)

Matrix products: default
BLAS/LAPACK: /projects/p31353/tjq4954/conda/daniel-sc-env/lib/libopenblasp-r0.3.25.so; LAPACK version 3.11.0

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 LC_MESSAGES=en_US.UTF-8 LC_PAPER=en_US.UTF-8
[8] LC_NAME=C LC_ADDRESS=C LC_TELEPHONE=C LC_MEASUREMENT=en_US.UTF-8 LC_IDENTIFICATION=C

time zone: UTC
tzcode source: system (glibc)

attached base packages:
[1] stats4 stats graphics grDevices utils datasets methods base

other attached packages:
[1] parallelly_1.37.1-9000 BSgenome.Hsapiens.UCSC.hg38_1.4.5 BSgenome_1.70.2 rtracklayer_1.62.0 BiocIO_1.12.0
[6] Biostrings_2.70.2 XVector_0.42.0 Azimuth_0.5.0 shinyBS_0.61.1 EnsDb.Hsapiens.v86_2.99.0
[11] ensembldb_2.26.0 AnnotationFilter_1.26.0 GenomicFeatures_1.54.3 AnnotationDbi_1.64.1 Biobase_2.62.0
[16] future_1.33.1 GenomicRanges_1.54.1 GenomeInfoDb_1.38.7 IRanges_2.36.0 S4Vectors_0.40.2
[21] BiocGenerics_0.48.1 Signac_1.12.0 pbmcref.SeuratData_1.0.0 SeuratData_0.2.2.9001 SeuratDisk_0.0.0.9021
[26] Seurat_5.0.2 SeuratObject_5.0.1 sp_2.1-3 HuMMuS_0.0.1

loaded via a namespace (and not attached):
[1] ProtGenerics_1.34.0 fs_1.6.3 matrixStats_1.2.0 spatstat.sparse_3.0-3 bitops_1.0-7 DirichletMultinomial_1.44.0
[7] devtools_2.4.5 TFBSTools_1.40.0 httr_1.4.7 RColorBrewer_1.1-3 profvis_0.3.8 tools_4.3.2
[13] sctransform_0.4.1 DT_0.32 utf8_1.2.4 R6_2.5.1 lazyeval_0.2.2 uwot_0.1.16
[19] rhdf5filters_1.14.1 urlchecker_1.0.1 withr_3.0.0 prettyunits_1.2.0 gridExtra_2.3 progressr_0.14.0
[25] cli_3.6.2 spatstat.explore_3.2-6 fastDummies_1.7.3 shinyjs_2.1.0 spatstat.data_3.0-4 readr_2.1.5
[31] ggridges_0.5.6 pbapply_1.7-2 Rsamtools_2.18.0 R.utils_2.12.3 sessioninfo_1.2.2 rstudioapi_0.15.0
[37] RSQLite_2.3.5 generics_0.1.3 gtools_3.9.5 ica_1.0-3 spatstat.random_3.2-3 googlesheets4_1.1.1
[43] dplyr_1.1.4 GO.db_3.18.0 Matrix_1.6-5 fansi_1.0.6 abind_1.4-5 R.methodsS3_1.8.2
[49] lifecycle_1.0.4 yaml_2.3.8 SummarizedExperiment_1.32.0 rhdf5_2.46.1 SparseArray_1.2.4 BiocFileCache_2.10.1
[55] Rtsne_0.17 grid_4.3.2 blob_1.2.4 promises_1.2.1 shinydashboard_0.7.2 crayon_1.5.2
[61] miniUI_0.1.1.1 lattice_0.22-5 cowplot_1.1.3 annotate_1.80.0 KEGGREST_1.42.0 pillar_1.9.0
[67] knitr_1.45 rjson_0.2.21 future.apply_1.11.1 codetools_0.2-19 fastmatch_1.1-4 leiden_0.4.3.1
[73] glue_1.7.0 data.table_1.15.2 remotes_2.4.2.1 vctrs_0.6.5 png_0.1-8 spam_2.10-0
[79] cellranger_1.1.0 gtable_0.3.4 poweRlaw_0.80.0 cachem_1.0.8 xfun_0.42 S4Arrays_1.2.1
[85] mime_0.12 pracma_2.4.4 survival_3.5-8 gargle_1.5.2 RcppRoll_0.3.0 ellipsis_0.3.2
[91] fitdistrplus_1.1-11 ROCR_1.0-11 nlme_3.1-164 usethis_2.2.3 bit64_4.0.5 filelock_1.0.3
[97] progress_1.2.3 RcppAnnoy_0.0.22 irlba_2.3.5.1 KernSmooth_2.23-22 colorspace_2.1-0 seqLogo_1.68.0
[103] DBI_1.2.2 tidyselect_1.2.1 curl_5.2.1 bit_4.0.5 compiler_4.3.2 hdf5r_1.3.10
[109] xml2_1.3.6 DelayedArray_0.28.0 plotly_4.10.4 scales_1.3.0 caTools_1.18.2 lmtest_0.9-40
[115] rappdirs_0.3.3 stringr_1.5.1 digest_0.6.35 goftest_1.2-3 presto_1.0.0 spatstat.utils_3.0-4
[121] rmarkdown_2.26 htmltools_0.5.7 pkgconfig_2.0.3 sparseMatrixStats_1.14.0 MatrixGenerics_1.14.0 dbplyr_2.4.0
[127] fastmap_1.1.1 rlang_1.1.3 htmlwidgets_1.6.4 shiny_1.8.0 zoo_1.8-12 jsonlite_1.8.8
[133] BiocParallel_1.36.0 R.oo_1.26.0 RCurl_1.98-1.14 magrittr_2.0.3 GenomeInfoDbData_1.2.11 dotCall64_1.1-1
[139] patchwork_1.2.0 Rhdf5lib_1.24.2 munsell_0.5.0 Rcpp_1.0.12 reticulate_1.35.0 stringi_1.8.3
[145] zlibbioc_1.48.0 MASS_7.3-60.0.1 plyr_1.8.9 pkgbuild_1.4.3 parallel_4.3.2 listenv_0.9.1
[151] ggrepel_0.9.5 deldir_2.0-4 CNEr_1.38.0 splines_4.3.2 tensor_1.5 hms_1.1.3
[157] igraph_2.0.2 spatstat.geom_3.2-9 RcppHNSW_0.6.0 reshape2_1.4.4 biomaRt_2.58.2 pkgload_1.3.4
[163] TFMPvalue_0.0.9 XML_3.99-0.16.1 evaluate_0.23 JASPAR2020_0.99.10 tzdb_0.4.0 httpuv_1.6.13
[169] RANN_2.6.1 tidyr_1.3.1 purrr_1.0.2 polyclip_1.10-6 scattermore_1.2 ggplot2_3.5.0
[175] xtable_1.8-4 restfulr_0.0.15 RSpectra_0.16-1 later_1.3.2 googledrive_2.1.1 viridisLite_0.4.2
[181] tibble_3.2.1 memoise_2.0.1 GenomicAlignments_1.38.2 cluster_2.1.6 globals_0.16.3

[r-trimbour]

Dear Daniel,

You can install the new version of HuMMuS with devtools::install_github("cantinilab/HuMMuS", ref="dev_SeuratV5").
It should solve your issue and issues with SeuratV5 in general 🙂

Let me know if it works !
Rémi