10X Xenium Starter Tutorial

Input Data

This tutorial uses SGE data generated with the 10x Genomics Xenium platform, and it has been cropped to a small region of the adult mouse brain for demonstration purposes.

File Format

10X Genomics Xenium platform outputs Spatial Gene Expression (SGE) data in a comma-separated values (CSV) format.

CSV File Format

"transcript_id","cell_id","overlaps_nucleus","feature_name","x_location","y_location","z_location","qv"
281827164036151,133793,1,"Sox10",2350.0232,4153.6846,16.592316,40.0
281827164036152,133793,1,"Sox10",2350.2585,4154.5225,17.237207,10.514394
281827164036164,151216,0,"Sox10",2350.5874,4277.699,14.285685,40.0

• "transcript_id": Unique identifier for each detected transcript molecule.
• "cell_id": ID of the segmented cell associated with the transcript.
• "overlaps_nucleus": 1 if the transcript overlaps the nucleus mask, 0 otherwise.
• "feature_name": Gene or other features name corresponding to the transcript.
• "x_location": X-coordinate of the transcript.
• "y_location": Y-coordinate of the transcript.
• "z_location": Z-coordinate (depth) of the transcript.
• "qv": Quality value indicating confidence in transcript detection.

Data Access

The example data is hosted on Zenedo ().

Follow the commands below to download the example data.

work_dir=/path/to/work/directory
cd $work_dir
wget  https://zenodo.org/records/15786632/files/xenium_starter.raw.tar.gz 
tar --strip-components=1 -zxvf xenium_starter.raw.tar.gz  

Set Up the Environment

Define paths to all required binaries and resources. Optionally, specify a fixed color map for consistent rendering.

# ====
# Replace each placeholder with the actual path on your system.  
# ====

work_dir=/path/to/work/directory        # path to work directory that contains the downloaded input data
cd $work_dir

# Define paths to required binaries and resources
spatula=/path/to/spatula/binary         # path to spatula executable
punkst=/path/to/punkst/binary           # path to FICTURE2/punkst executable
tippecanoe=/path/to/tippecanoe/binary   # path to tippecanoe executable
pmtiles=/path/to/pmtiles/binary         # path to pmtiles executable
aws=/path/to/aws/cli/binary             # path to AWS CLI binary

# (Optional) Define path to color map. 
cmap=/path/to/color/map                 # Path to the fixed color map for rendering. cartloader provides a fixed color map at cartloader/assets/fixed_color_map_256.tsv.

# Number of jobs
n_jobs=10                               # If not specify, the number of jobs defaults to 1.

# Activate the bioconda environment
conda activate ENV_NAME                 # replace BIOENV_NAME with your bioconda environment name

Define data ID and analysis parameters:

# Unique identifier for your dataset
DATA_ID="xenium_hippo"                  # change this to reflect your dataset name
PLATFORM="10x_xenium"                   # platform information
SCALE=1                                 # coordinate to micrometer scaling factor

# LDA parameters
train_width=12                           # define LDA training hexagon width (comma-separated if multiple widths are applied)
n_factor=6,12                            # define number of factors in LDA training (comma-separated if multiple n-factor are applied)

How to Define Scaling Factors for Xenium

The Xenium example data currently used here provides SGE in micrometer units. Use define scaling factor from coordinate to micrometer as 1.

SGE Format Conversion

Convert the raw input to the unified SGE format. See more details in Reference page.

cartloader sge_convert \
  --makefn sge_convert.mk \
  --platform ${PLATFORM} \
  --in-csv ./input.tsv.gz \
  --units-per-um ${SCALE} \
  --out-dir ./sge \
  --exclude-feature-regex '^(BLANK|NegCon|NegPrb)' \
  --sge-visual \
  --spatula ${spatula} \
  --n-jobs ${n_jobs}

Parameter	Required	Type	Description
--platform	required	string	Platform (options: "10x_visium_hd", "seqscope", "10x_xenium", "bgi_stereoseq", "cosmx_smi", "vizgen_merscope", "pixel_seq", "generic")
--in-csv	required	string	Path to the input TSV/CSV file
--units-per-um	required	float	Scale to convert coordinates to microns (default: 1.0)
--out-dir	required	string	Output directory for the converted SGE files
--makefn		string	File name for the generated Makefile (default: sge_convert.mk)
--exclude-feature-regex		regex	Pattern to exclude control features
--sge-visual		flag	Enable SGE visualization step (generates diagnostic image) (default: FALSE)
--spatula		string	Path to the spatula binary (default: spatula)
--n-jobs		int	Number of parallel jobs for processing (default: 1)

FICTURE Analysis

Compute spatial factors using punkst (FICTURE2 mode). See more details in Reference page.

cartloader run_ficture2 \
  --makefn run_ficture2.mk \
  --main \
  --in-transcript ./sge/transcripts.unsorted.tsv.gz \
  --in-feature ./sge/feature.clean.tsv.gz \
  --in-minmax ./sge/coordinate_minmax.tsv \
  --cmap-file ${cmap} \
  --exclude-feature-regex '^(mt-.*$|Gm\d+$)' \
  --out-dir ./ficture2 \
  --width ${train_width} \
  --n-factor ${n_factor} \
  --spatula ${spatula} \
  --ficture2 ${punkst} \
  --n-jobs ${n_jobs} \
  --threads ${n_jobs}

Parameter	Required	Type	Description
--main	required 1	flag	Enable cartloader to run all five steps
--in-transcript	required	string	Path to input transcript-level SGE file
--out-dir	required	string	Path to output directory
--width	required	int or comma-separated list	LDA training hexagon width(s)
--n-factor	required	int or comma-separated list	Number of LDA factors
--makefn		string	File name for the generated Makefile (default: run_ficture2.mk )
--in-feature		string	Path to input feature file
--in-minmax		string	Path to input coordinate min/max file
--cmap-file		string	Path to color map file
--exclude-feature-regex		regex	Pattern to exclude features
--spatula		string	Path to the spatula binary (default: spatula)
--ficture2		string	Path to the punkst directory (defaults to punkst repository within submodules directory of cartloader)
--n-jobs		int	Number of parallel jobs (default: 1)
--threads		int	Number of threads per job (default: 1)

_{¹: cartloader requires the user to specify at least one action. Available actions includes: --tile to run tiling step; --segment to run segmentation step; --init-lda to run LDA training step; --decode to run decoding step; --summary to run summarization step; --main to run all above five actions.}

cartloader Compilation

Generate pmtiles and web-compatible tile directories. See more details in Reference page.

cartloader run_cartload2 \
  --makefn run_cartload2.mk \
  --fic-dir ./ficture2 \
  --out-dir ./cartload2 \
  --id ${DATA_ID} \
  --spatula ${spatula} \
  --pmtiles ${pmtiles} \
  --tippecanoe ${tippecanoe} \
  --n-jobs ${n_jobs} \
  --threads ${n_jobs}

Parameter	Required	Type	Description
--fic-dir	required	string	Path to the input directory containing FICTURE2 output
--out-dir	required	string	Path to the output directory for PMTiles and web tiles
--id	required	string	Dataset ID used for naming outputs and metadata
--makefn		string	File name for the generated Makefile (default: run_cartload2.mk)
--spatula		string	Path to the spatula binary (default: spatula)
--pmtiles		string	Path to the pmtiles binary (default: pmtiles)
--tippecanoe		string	Path to the tippecanoe binary (default: tippecanoe)
--n-jobs		int	Number of parallel jobs (default: 1)
--threads		int	Number of threads per job (default: 1)

Upload to Data Repository

Choose a data repository to host/share your output

cartloader supports two upload options (AWS and Zenodo) for storing PMTiles of SGE and spatial factors in a data repository.

Choose the one that best suits your needs.

AWS Uploads

Upload the generated cartloader outputs to your designated AWS S3 directory:

# AWS S3 target location for cartostore
AWS_BUCKET="EXAMPLE_AWS_BUCKET"         # replace EXAMPLE_AWS_BUCKET with your actual S3 bucket name

cartloader upload_aws \
  --in-dir ./cartload2 \
  --s3-dir "s3://${AWS_BUCKET}/${DATA_ID}" \
  --aws ${aws} \
  --n-jobs ${n_jobs}

Parameter	Required	Type	Description
--in-dir	required	string	Path to the input directory containing the cartloader compilation output
--s3-dir	required	string	Path to the target S3 directory for uploading
--aws		string	Path to the AWS CLI binary
--n-jobs		int	Number of parallel jobs

Zenodo Uploads

Upload the generated cartloader outputs to your designated Zenodo deposition or a new deposition.

zenodo_token=/path/to/zenodo/token/file    # replace /path/to/zenodo/token/file by path to your zenodo token file

cartloader upload_zenodo \
  --in-dir ./cartload2 \
  --upload-method catalog \
  --zenodo-token $zenodo_token \
  --create-new-deposition \
  --title  "Yur Title" \
  --creators "Your Name" \
  --description "This is an example description"

Parameter	Required	Type	Description
--in-dir	required	string	Path to the input directory containing the cartloader compilation output
--upload-method	required	string	Method to determine which files to upload. Options: all to upload all files in --in-dir; catalog to upload files listed in a catalog YAML file, user_list to upload files explicitly listed via --in-list
--catalog-yaml		string	Required if --upload-method catalog. Path to the catalog.yaml file generated in run_cartload2. If absent, will use the catalog.yaml in the input directory specified by --in-dir.
--zenodo-token	required	string	Path to your Zenodo access file
--create-new-deposition		flag	a new Zenodo deposition will be created.
--title	required	string	Required if --create-new-deposition. Title for the new Zenodo deposition.
--creators	required	list of str	List of creators in "Lastname, Firstname" format.

Output Data

See more details of output at the Reference pages for run_ficture2 and run_cartload2.

Spatial Factor Inference from FICTURE

Below is an example of spatial factor inference results produced by FICTURE using a training width of 12, 12 factors, a fit width of 12, and an anchor resolution of 6.

Factor	RGB	Weight	PostUMI	TopGene_pval	TopGene_fc	TopGene_weight
2	101,255,101	0.16118	699466	Cabp7,Gfap,Arc,Orai2,2010300C02Rik,Arhgap12,Sipa1l3,Kctd12,Lyz2,Shisa6,Pou3f1,Rmst,Arhgap6,Clmn,Chrm2,Cntn6,Cux2,Fibcd1,Wfs1,Pkib	Cabp7,Orai2,Slc6a3,Gfap,Arc,Th,Arhgap6,Arhgap12,Pkib,Lyz2,Pou3f1,Ror1,Shisa6,Rmst,Sipa1l3,Vip,Chrm2,Unc13c,2010300C02Rik,Ebf3	Cabp7,Gfap,Slc17a7,Arc,2010300C02Rik,Sipa1l3,Kctd12,Arhgap12,Clmn,Orai2,Acsbg1,Shisa6,Bcl11b,Igfbp5,Bhlhe22,Pou3f1,Rmst,Lyz2,Ntsr2,Wfs1
0	255,101,101	0.13961	605851	Prox1,Calb1,Bhlhe22,Plekha2,Rasgrf2,Pdzd2,Pip5k1b,Rasl10a,Prdm8,Epha4,2010300C02Rik,Npnt,Slc44a5,Nrn1,Bcl11b,Pdyn,Thsd7a,Dpyd,Cntnap5b,Zfpm2	Pdyn,Npnt,Prox1,Plekha2,Calb1,Pdzd2,Rasgrf2,Cntnap5b,Rasl10a,Slc44a5,Bhlhe22,Dpyd,Prdm8,Zfpm2,Thsd7a,Pip5k1b,Syt17,Gsg1l,Npy2r,Cdh9	Prox1,Nrn1,Epha4,2010300C02Rik,Slc17a7,Bhlhe22,Pip5k1b,Calb1,Bcl11b,Rasgrf2,Cpne6,Plekha2,Cpne4,Prdm8,Igfbp5,Rasl10a,Bhlhe40,Tanc1,Gad1,Pdzd2
1	237,238,0	0.12496	542277	Cpne4,Neurod6,Slc17a7,Cpne6,Epha4,Slit2,Nrn1,Prdm8,Sema3e,Dkk3,Nwd2,Npy2r,Strip2,Bdnf,Plcxd2,Shisa6,Tanc1,Rspo2,Mdga1,Car4	Slit2,Cpne4,Neurod6,Rspo2,Npy2r,Strip2,Prss35,Cpne6,Stard5,Bdnf,Prdm8,Nwd2,Sema3e,Plcxd2,Epha4,Shisa6,Slc17a7,Trp73,Nrn1,Tanc1	Slc17a7,Neurod6,Epha4,Nrn1,Cpne4,Cpne6,Cabp7,Dkk3,Dner,Slit2,Tanc1,Prdm8,Sema3e,Sipa1l3,2010300C02Rik,Shisa6,Nwd2,Bhlhe22,Car4,Plcxd2
4	101,101,255	0.10988	476853	Garnl3,Hs3st2,Nxph3,Myl4,Cplx3,Rprm,Satb2,Dpy19l1,Fhod3,Igfbp6,Tle4,Dkk3,Grik3,Ccn2,Gfra2,Gadd45a,Arc,Slc17a7,Igfbp4,Igsf21	Hs3st2,Trbc2,Myl4,Cplx3,Nxph3,Rxfp1,Igfbp6,Satb2,Garnl3,Rprm,Ccn2,Arhgap25,Gadd45a,Fhod3,Sema3a,Sdk2,Foxp2,Tox,Sla,Igsf21	Slc17a7,Garnl3,Dkk3,Nrn1,Arc,Dpy19l1,Neurod6,Id2,Tle4,Igfbp4,Rprm,Nxph3,Gfra2,Epha4,Hs3st2,Myl4,Grik3,2010300C02Rik,Satb2,Fhod3
3	101,254,255	0.1069	463920	Wfs1,Neurod6,Pou3f1,Epha4,Sema3e,Fibcd1,Dkk3,Sorcs3,Igfbp4,Bcl11b,Cpne8,2010300C02Rik,Trpc4,Nell1,Cpne6,Ndst4,Ndst3,Arhgap12,Penk,Mdga1	Wfs1,Pou3f1,Fibcd1,Ndst4,Sorcs3,Cpne8,Penk,Sema3e,Ndst3,Igfbp4,Trpc4,Neurod6,Nell1,Htr1f,Strip2,Bcl11b,Ror1,Mdga1,Dkk3,Epha4	Neurod6,Epha4,Slc17a7,2010300C02Rik,Dkk3,Nrn1,Cpne6,Wfs1,Bcl11b,Pou3f1,Sipa1l3,Sema3e,Arc,Dner,Igfbp4,Pip5k1b,Arhgap12,Cabp7,Sorcs3,Trpc4
5	255,101,254	0.09002	390640	Aqp4,Acsbg1,Ntsr2,Id2,Slc39a12,Rfx4,Rorb,Gfap,Hapln1,Gli3,Rmst,Cdh20,Igfbp5,Clmn,Ppp1r1b,Mapk4,Bhlhe40,Angpt1,Cd44,Gng12	Aqp4,Slc39a12,Rfx4,Ntsr2,Acsbg1,Id2,Gli3,Hapln1,Cd44,Rorb,Cdh20,Eya4,Rmst,Gjb2,Prph,Angpt1,Ppp1r1b,Pdzrn3,Cdh4,Gfap	Acsbg1,Id2,Aqp4,Ntsr2,Gfap,Slc39a12,Rorb,Clmn,Igfbp5,Dkk3,Dner,Bhlhe40,Rfx4,Rmst,Hapln1,Gng12,Mapk4,Cdh20,Tle4,Ppp1r1b
6	255,178,101	0.06358	275899	Cldn5,Ly6a,Fn1,Adgrl4,Pecam1,Igf2,Laptm5,Kdr,Trem2,Acvrl1,Dcn,Siglech,Emcn,Cd53,Cd93,Sox17,Aldh1a2,Nostrin,Pglyrp1,Paqr5	Cldn5,Cd93,Acvrl1,Fn1,Nostrin,Slfn5,Pecam1,Sox17,Adgrl4,Emcn,Aldh1a2,Igf2,Ly6a,Paqr5,Fmod,Dcn,Kdr,Zfp366,Fgd5,Spp1	Ly6a,Cldn5,Fn1,Laptm5,Adgrl4,Igf2,Pecam1,Kctd12,Car4,Siglech,Gfap,Trem2,Kdr,Dcn,Cd53,Lyz2,Acvrl1,Nr2f2,Cobll1,Emcn
7	178,255,101	0.06088	264215	Gad2,Gad1,Pvalb,Rab3b,Dner,Sst,Kcnmb2,Btbd11,Cntnap4,Cdh13,Cort,Hapln1,Igf1,Col19a1,Zfp536,Rasgrf2,Necab1,Myo16,Cacna2d2,Lypd6	Sst,Gad2,Cort,Rab3b,Kcnmb2,Pvalb,Gad1,Crh,Igf1,Col19a1,Btbd11,Cntnap4,Rbp4,Pde11a,Pthlh,Cbln4,Cdh13,Hapln1,Syt2,Cacna2d2	Gad1,Gad2,Dner,Pvalb,Rab3b,Id2,Cdh13,Zfp536,Rasgrf2,Hapln1,Kctd12,Necab1,Cntnap4,Rims3,Igfbp4,Kcnmb2,Btbd11,Neto2,Plcxd3,Lypd6
8	0,223,95	0.05087	220765	Gjc3,Opalin,Sox10,Sema6a,Dpy19l1,Zfp536,Adamtsl1,Tmem163,Gng12,Clmn,Cobll1,Cdh20,Vwc2l,Gpr17,Prox1,Fign,Adamts2,Sema3d,Arhgef28,Chrm2	Opalin,Sox10,Gjc3,Adamtsl1,Sema6a,Tmem163,Zfp536,Fign,Adamts2,Cobll1,Dpy19l1,Gpr17,Vwc2l,Cdh20,Gng12,Sema3d,Clmn,Carmn,Chrm2,Parm1	Gjc3,Opalin,Sox10,Dpy19l1,Clmn,Sema6a,Gng12,Zfp536,Prox1,Gfap,Tmem163,Cdh20,Arhgef28,Cobll1,Bhlhe40,Adamtsl1,Igfbp5,Vwc2l,Chrm2,Tanc1
9	101,178,255	0.03417	148267	Slc17a6,Rims3,Necab1,Calb2,Rnf152,Nell1,Clmn,Tmem163,Pde7b,Rorb,Cux2,Foxp2,Opn3,Unc13c,Inpp4b,Btbd11,Thsd7a,Fign,Sema6a,Nrn1	Slc17a6,Rims3,Necab1,Opn3,Calb2,Rnf152,Fign,Unc13c,Nell1,Tmem163,Btbd11,Inpp4b,Foxp2,Cux2,Pde7b,Plch1,Deptor,Tox,Rorb,Thsd7a	Rims3,Slc17a6,Nrn1,Necab1,Clmn,Calb2,Epha4,Nell1,Dner,Slc17a7,Rorb,Rnf152,Prox1,Bhlhe40,Pde7b,Nrep,Thsd7a,Tmem163,Tanc1,Sema6a
10	178,101,255	0.029	125838	Nwd2,Calb2,Syt6,Necab2,Sncg,Vat1l,Slc17a6,Gucy1a1,Kctd8,Nrp2,Tmem255a,Chat,Tacr1,Cacna2d2,Thsd7a,Ano1,Dner,Cbln1,Cntnap4,Cbln4	Chat,Sncg,Tacr1,Kctd8,Vat1l,Nwd2,Calb2,Syt6,Cbln1,Ano1,Necab2,Gucy1a1,Slc17a6,Tmem255a,Cacna2d2,Cbln4,Nrp2,Th,Pde11a,Sema3d	Nwd2,Nrn1,Calb2,Necab2,Dner,Syt6,Slc17a6,Cpne4,Gucy1a1,Nrp2,Vat1l,Thsd7a,Sncg,Slc17a7,Kctd8,Cacna2d2,Rmst,Garnl3,Tmem255a,Cntnap4
11	255,153,204	0.02895	125631	Pdgfra,Cspg4,Gpr17,Cd24a,Sox11,Sema3d,Spag16,Igfbp5,Gjc3,Trp73,Gfap,Meis2,Nrep,Sox10,Sema5b,Cobll1,Rmst,Vwc2l,Tmem255a,Hat1	Pdgfra,Cspg4,Gpr17,Spag16,Cd24a,Sema3d,Trp73,Sox11,Chodl,Sema5b,Ebf3,Prph,Tmem255a,Plch1,Cdh4,Slc6a3,Sox10,Gjc3,Vwc2l,Cobll1	Pdgfra,Gfap,Igfbp5,Gpr17,Cspg4,Gjc3,Sox11,Cd24a,Nrep,Meis2,Dner,Dkk3,Sox10,Rmst,Bhlhe40,Gng12,Dpy19l1,Sema3d,Sema6a,Cobll1

Packed SGE and Spatial Factor Outputs from run_cartload2

The packed SGE data and spatial factor inferences generated by FICTURE are available in PMTile format on Zenodo: DOI:10.5281/zenodo.15824967.

These datasets can also be loaded directly using the following catalog YAML file:
https://zenodo.org/records/15824968/catalog.yaml