Skip to content

README.md

Latest commit

 

History

History
230 lines (184 loc) · 7.08 KB

README.md

File metadata and controls

230 lines (184 loc) · 7.08 KB

SpatialData.data

for class infrastructure, see SpatialData

for visualization capabilites, see SpatialData.plot

Introduction

Spatialdata.data package provides utilities for accessing, reading and generating SpatialData datasets.

scverse SpatialData examples are available through

  1. Bioc's NSF OSN bucket and
  2. scverse's spatialdata-sandbox (https://spatialdata.scverse.org/en/latest/tutorials/notebooks/datasets/README.html)

which are accessible from within R, using BiocFileCache.

SpatialData.data uses basilisk to interface and maintain multiple versions of python's spatialdata package (0.3.0, 0.5.0 and 0.7.2) for reading and writing to .zarr packages.

basilisk environments (only 0.5.0 and 0.7.2) are also accompanied by the dummy-spatialdata package that generates toy spatialdata examples whose elements are customized by the user.

Installation

BiocManager::install("HelenaLC/SpatialData")
BiocManager::install("HelenaLC/SpatialData.data")

To interrogate our S3 bucket you will need paws installed; it is not necessary for retrievals.

Introduction

Zipped .zarr archives

Data from a variety of technologies has been made available as SpatialData .zarr stores here. These, in turn, have been deposited in Bioconductor's NSF Open Storage Network also and can be retrieved with caching support using r BiocStyle::Biocpkg("BiocFileCache").

Any spatialdata dataset can be retrieved (once) into some location, and read into R. We use dataset-specific functions, or load_data:

(x <- load_data("ColorectalCarcinomaMIBITOF")) # stub can be used
class: SpatialData
- images(3):
  - point16_image (3,1024,1024)
  - point23_image (3,1024,1024)
  - point8_image (3,1024,1024)
- labels(3):
  - point16_labels (1024,1024)
  - point23_labels (1024,1024)
  - point8_labels (1024,1024)
- points(0):
- shapes(0):
- tables(1):
  - table (36,3309)
coordinate systems:
- point16(2): point16_image point16_labels
- point23(2): point23_image point23_labels
- point8(2): point8_image point8_labels

or

# from biocOSN
x <- ColorectalCarcinomaMIBITOF()

# from sandbox (Zarr v3)
x <- ColorectalCarcinomaMIBITOF(source = bucket_path("sandbox"))

We can check all available datasets below:

SpatialData.data_list()
                       Function             Technology       S3_buckets                           Format
1         MouseIntestineVisHD()              Visium HD biocOSN, sandbox 0.3.0 (Zarr v2), 0.7.2 (Zarr v3)
2             MouseBrainVisHD()              Visium HD          sandbox                  0.3.0 (Zarr v2)
3               MouseBrainVis()                 Visium          sandbox                  0.7.2 (Zarr v3)
4   LungAdenocarcinomaMCMICRO() CyCIF (MCMICRO output)          biocOSN                  0.3.0 (Zarr v2)
5           MouseBrainMERFISH()                MERFISH biocOSN, sandbox 0.3.0 (Zarr v2), 0.7.2 (Zarr v3)
6           MouseLiverMERFISH()                MERFISH          sandbox                  0.7.2 (Zarr v3)
7  ColorectalCarcinomaMIBITOF()               MIBI-TOF biocOSN, sandbox 0.3.0 (Zarr v2), 0.7.2 (Zarr v3)
8        MulticancerSteinbock() IMC (Steinbock output)          biocOSN                  0.3.0 (Zarr v2)
9     JanesickBreastVisiumEnh()                 Visium biocOSN, sandbox 0.3.0 (Zarr v2), 0.7.2 (Zarr v3)
10   JanesickBreastXeniumRep1()                 Xenium biocOSN, sandbox 0.3.0 (Zarr v2), 0.7.2 (Zarr v3)
11   JanesickBreastXeniumRep2()                 Xenium          biocOSN                  0.3.0 (Zarr v2)
12         HumanLungMulti_10x()                 Xenium biocOSN, sandbox 0.3.0 (Zarr v2), 0.7.2 (Zarr v3)
13             Breast2fov_10x()       Xenium (trimmed)   biocOSN_Xenium 0.3.0 (Zarr v2), 0.7.2 (Zarr v3)
14               Lung2fov_10x()       Xenium (trimmed)   biocOSN_Xenium 0.3.0 (Zarr v2), 0.7.2 (Zarr v3)
15           SpaceMHelaniH3T3()                 SpaceM          sandbox                  0.7.2 (Zarr v3)
```

To interrogate the bucket for available (zipped) .zarr archives:

Sys.setenv(AWS_REGION = "us-east-1")
if (requireNamespace("paws")) available("biocOSN")

SpatialData-io

SpatialData.data also provides access to some raw spatial omic readouts. These data bundle can then be converted into SpatialData objects using spatialdata-io pyton package via basilisk.

available("biocOSN_Xenium")
checking Bioconductor OSN bucket (Xenium readouts) ...
[1] "README.html"                                         "Xenium_Prime_MultiCellSeg_Mouse_Ileum_tiny_outs.zip"
[3] "Xenium_V1_human_Breast_2fov_outs.zip"                "Xenium_V1_human_Lung_2fov_outs.zip"

We use basilisk to convert these readouts into various SpatialData formats:

  • 0.3.0 (Zarr v2),
  • 0.5.0 (Zarr v2) and
  • 0.7.2 (Zarr v3)

We use options(sd_version) to set the SpatialData version.

options(sd_version = "0.3.0")
(x <- Breast2fov_10x(source = bucket_path("biocOSN_Xenium")))

Using spatialdata version 0.3.0
class: SpatialData
- images(1):
  - morphology_focus (5,3529,5792)
- labels(2):
  - cell_labels (3529,5792)
  - nucleus_labels (3529,5792)
- points(1):
  - transcripts (1113950)
- shapes(3):
  - cell_boundaries (7275,circle)
  - cell_circles (7275,circle)
  - nucleus_boundaries (7020,circle)
- tables(1):
  - table (280,7275)
coordinate systems:
- global(7): morphology_focus cell_labels ... nucleus_boundaries transcripts

Dummy SpatialData objects

SpatialData.data package incorporates the dummy-spatialdata python package from PyPI (https://pypi.org/project/dummy-spatialdata/) via basilisk to generate toy spatialdata objects in multiple spatialdata versions.

zarrfile <- tempfile(fileext = ".zarr")
generate_dataset(
  file = zarrfile, 
  sd_version = "0.5.0",
  images = list(
    list(type = "rgb", n_layers = 4L, coordinate_system="global"),
    list(type = "grayscale", n_layers = 1L, coordinate_system="global")
  ),
  shapes = list(
    list(n_shapes=12L, coordinate_system="global")
  ),
  points = list(
    list(n_points=12L)
  ),
  coordinate_systems = list(
    global = list(
      transformations = list("affine"), 
      shape = list(x=2000L, y=2000L)
    )
  )
)
Using spatialdata version 0.5.0
[1] "/var/folders/vf/d8kg507x41xfh6z9vgv9skksdsn29w/T//RtmpXl8ziv/file11cae4154ccb2.zarr"

sd <- SpatialData::readSpatialData(zarrfile)
sd
class: SpatialData
- images(2):
  - image_0 (3,2000,2000)
  - image_1 (1,2000,2000)
- labels(0):
- points(1):
  - point_0 (12)
- shapes(1):
  - shape_0 (12,polygon)
- tables(0):
coordinate systems:
- global(3): image_0 image_1 shape_0

image(sd, 1)
class:  ImageArray (MultiScale) 
Scales (4): (3,2000,2000) (3,1000,1000) (3,500,500) (3,250,250)