METASPACE to S2IsoMEr

Bishoy Wadie

In this vignette, we’ll demonstrate how to use the METASPACE converter to transform METASPACE datasets into a format compatible with S2IsoMEr. Note that these datasets will be in a single-pixel format rather than single-cell unless an additional mapping from a cell-segmentation output links pixels to cells.

Since the METASPACE converter is written in Python, we’ll provide a snippet demonstrating how to first export a METASPACE dataset as an AnnData object and then import it into R using the anndataR package. This vignette provides only a basic overview of the METASPACE converter. For installation instructions and more detailed usage, please refer to the documentation.

We’ll use the METASPACE datasets from the same NASH model referenced in the ORA vignette. For more details about the dataset, refer to that vignette.

Loading packages

require(S2IsoMEr)
require(dplyr)
require(tidyr)
require(Seurat)
require(anndataR)

Getting Dataset IDs

As in the ORA vignette, we’ll start by downloading the datasets and mapping the condition to the dataset ID.

NASH_scm_tmp = tempfile()
download.file("https://zenodo.org/records/13318721/files/NASH_scm_dataset.rds", destfile = NASH_scm_tmp)
NASH_scm = readRDS(NASH_scm_tmp)

condition_metadata_tmp = tempfile()
download.file("https://zenodo.org/records/13318721/files/spacem_scm_matrices.rds", destfile = condition_metadata_tmp)
condition_metadata = readRDS(condition_metadata_tmp)[["metaspace_dataset_names"]]
condition_metadata$dataset_name = sub(".ibd.*", "", condition_metadata$dataset_name)

metaspace_annotations_tmp = tempfile()
download.file("https://zenodo.org/records/13318721/files/SpaceM_metaspace_ds_annotations.rds", destfile = metaspace_annotations_tmp)
metaspace_annotations = readRDS(metaspace_annotations_tmp)

Get metadata

scm = NASH_scm$scm %>% 
  as.matrix() %>% 
  t()

conds = NASH_scm$metadata %>%
  column_to_rownames("Cell")
conds = conds[colnames(scm),]

conds_unique = conds %>% 
  dplyr::distinct()

metaspace_annotations = metaspace_annotations %>% 
  dplyr::left_join(condition_metadata, by = c("ds_name" = "dataset_name")) %>%
  dplyr::rename("Replicate" = "Condition") %>%
  dplyr::left_join(conds_unique) %>% 
  dplyr::select(ds_id, Replicate, Condition) %>% 
  dplyr::distinct()

Filter conditions

cond_x = "U"
cond_y = "F"

Dss_id_conds = metaspace_annotations %>% 
  dplyr::filter(Condition %in% c(cond_x, cond_y))

#Change "path/to/dir" with your local directory
write.csv(Dss_id_conds, "path/to/dir/Ds_ids_for_METASPACE_converter.csv", row.names = F)

METASPACE to AnnData

Now you can use the exported metadata with dataset IDs and use METASPACE converter to get the merged AnnData object. You’ll need to run the following python chunk in your own python environment to export the AnnData object to an H5AD file, which you’ll then import back into R in the next step.

from metaspace_converter import metaspace_to_anndata
import scanpy as sc
import pandas as pd

#Change /path/to/dir with local path
dss = pd.read_csv("/path/to/dir/Ds_ids_for_METASPACE_converter.csv")

dss_list = list(dss['ds_id'])

FDR_level = 0.1
DB = ("HMDB", "v4")
Adatas = []

for ds in dss_list:
    adata = metaspace_to_anndata(
        dataset_id=ds,
        fdr=FDR_level,
        database=DB)
    Replicate = dss['Replicate'][dss['ds_id'] == ds].values[0]
    adata.obs.index = Replicate + "_" + adata.obs.index
    adata.obs['Condition'] = dss['Condition'][dss['ds_id'] == ds].values[0]
    adata.obs['Replicate'] = Replicate
    Adatas.append(adata)

merged = sc.concat(Adatas, axis=0, join='outer', merge='first')
merged.var["offSample"] = merged.var["offSample"].astype(str)

#Change /path/to/dir with local path
merged.write("/path/to/dir/NASH_FU_Anndata_FDR_0.1_HMDB_v4.h5ad")

Read AnnData into R

adata = anndataR::read_h5ad("/path/to/dir/NASH_FU_Anndata_FDR_0.1_HMDB_v4.h5ad")
seurat_obj = anndataR::to_Seurat(adata)

scm = seurat_obj@assays$RNA@data %>% as.matrix()
grps =  seurat_obj@meta.data$Condition %>% as.character()

Initialize S2IsoMEr object for ORA

Now that we have the input single-pixel matrix and the corresponding condition for each pixel, we can create the S2IsoMEr object for ORA enrichment as before.

cond_x = "U"
cond_y = "F"

ORA_boot_obj = initEnrichment(scmatrix = scm, conditions = grps,
                     enrichment_type = "ORA",annot_db = "HMDB",
                     consider_isomers = T, consider_isobars = T,
                     polarization_mode = "positive",
                     background_type = "sub_class",
                     molecule_type = "Metabo",
                     condition.x = cond_x,
                     condition.y = cond_y)