CosMX SMI Starter Tutorial¶
Input Data¶
The input data is from an adult mouse hippocampus, extracted by masking a coronal brain section. The original full-section
File Format
The CosMx SMI by NanoString generates high-resolution spatial transcriptomics data with single-molecule resolution with a comma-separated values (CSV) table.
CSV File Format
1 2 3 4 |
|
fov
: The field of view (FOV) number.cell_ID
: Unique identifier for a single cell within a given FOV. 0 if background or unassigned molecules.x_global_px
,y_global_px
: Global pixel coordinates relative to the tisse.x_local_px
,y_local_px
: The x or y position (in pixels) relative to the given FOV.z
: Z-plane index representing the depth (optical section) where the transcript was detected.target
: Name of the target.CellComp
: Subcellular location of target.
Data Access
The example data is hosted on Zenedo ().
Follow the commands below to download the example data.
1 2 3 4 |
|
Set Up the Environment¶
Define paths to all required binaries and resources. Optionally, specify a fixed color map for consistent rendering.
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 |
|
Define data ID and analysis parameters:
1 2 3 4 5 6 7 8 |
|
How to Define Scaling Factors for CosMX SMI?
According to the README.html provided with the Pixel-seq dataset, each pixel has an edge length of 120 nm. To calculate the number of pixels per micrometer, use the formula: scale = 1000 / 120.
SGE Format Conversion¶
Convert the raw input to the unified SGE format. See more details in Reference page.
1 2 3 4 5 6 7 8 9 10 |
|
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.
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 |
|
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.
1 2 3 4 5 6 7 8 9 10 |
|
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:
1 2 3 4 5 6 7 8 |
|
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.
1 2 3 4 5 6 7 8 9 10 |
|
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 |
---|---|---|---|---|---|---|
1 | 237,238,0 | 0.17207 | 548469625 | Tmsb4x,Calm2,Snap25,Calm1,Ywhaz,Slc17a7,Cck,Dnm1,Ywhag,Aldoa,Hsp90ab1,Calm3,Atp2b1,Nell2,Rtn3,Snca,Gpm6a,Prkcb,Prkca,Cyfip2 | Cck,Camk4,Prkca,Camk1d,Slc17a7,Ywhag,Epha6,Atp2b1,Calm2,Sv2b,Snca,Syn2,Prkcb,Lpl,Nell2,Snap25,Stmn2,Pde1a,Kcnq5,Cnksr2 | Tmsb4x,Snap25,Calm2,Calm1,Ywhaz,Atp1b1,Aldoa,Hsp90ab1,Dnm1,Slc17a7,Rtn3,Calm3,Camk2a,Ywhag,Gpm6a,Cck,Ppp3ca,Rtn1,Nrgn,Cpe |
0 | 255,101,101 | 0.16 | 510000383 | Camk2b,Adcy1,Ppp3ca,Kalrn,Nrgn,Olfm1,Ppfia2,Ctxn1,Gria1,Gria2,Prkce,Ryr2,Camk2a,Grin2a,Gabrb3,Zbtb20,Wasf1,Tiam1,Ntrk3,Mapk1 | Ppfia2,Adcy1,Sema5a,Camk2b,Kalrn,Plxna4,Calb1,Tiam1,Jun,Rasgrf2,Nedd4l,Dapk1,Grin2a,Npy1r,Itga8,Ryr2,Auts2,Ppp3ca,Dgkg,Grm7 | Malat1,Camk2b,Camk2a,Calm1,Ppp3ca,Tmsb4x,Adcy1,Nrgn,Calm2,Olfm1,Atp1b1,Ndrg4,Gria2,Ywhaz,Kalrn,Gria1,Ctxn1,Prkce,Aldoa,Snap25 |
3 | 101,254,255 | 0.11601 | 369772892 | Malat1,Hnrnpa2b1,Tardbp,Tnik,Grk3,Mid1,Ptprs,Nf1,Prkn,Vegfa,Kidins220,Tnrc6a,Dmd,Camk2g,Uggt2,Pou2f1,Meg3,Sema6d,Spag9,Ntsr2 | Gal,Malat1,Avp,Cd109,Grk3,Lilra5,C2,Prkn,Crhr1,Ngfr,Gcgr,Dmd,Vegfa,C3,Dapk2,Pcsk6,Nrf1,Uggt2,P2rx4,Slc18a2 | Malat1,Hnrnpa2b1,Hmgb1,Meg3,Gria2,Ptprs,Tardbp,Gnao1,Tnik,Apoe,Kidins220,Slc1a2,Ahcyl1,Nf1,Bcan,Mid1,Spag9,Camk2g,Fus,Tuba1a/b/c |
5 | 255,101,254 | 0.09996 | 318610467 | Nap1l5,Atp1b1,Gad1,Ndrg4,Meg3,Cnr1,Sncb,Gad2,Sst,Snap25,Syt1,Zwint,Tmsb10,Snrpn,Gnas,Npy,Sv2a,Cacnb4,Ldhb,Mdh1 | Sst,Gad2,Npy,Cnr1,Nap1l5,Gad1,Pvalb,Slc32a1,Vip,Gap43,Adgrl2,Cntn4,Sncb,Atp2b4,Cckbr,Grik1,Grm8,Rims2,Cacnb4,Tmsb10 | Atp1b1,Malat1,Snap25,Ndrg4,Meg3,Nap1l5,Ckb,Rtn3,Snrpn,Gnas,Calm1,Syt1,Zwint,Mdh1,Dnm1,Calm2,Aldoa,Sncb,Gad1,Dynll2 |
2 | 101,255,101 | 0.09064 | 288906127 | Plp1,Ptgds,Scd2,Apod,Mag,Cryab,Ndrg1,Mog,Ugt8a,Bin1,Gpr37,Aspa,Mobp,Fa2h,Gsn,Mbp,Abca2,Cntn2,Slc44a1,Jam3 | Plp1,Mag,Mog,Ugt8a,Aspa,Ptgds,Cryab,Apod,Fa2h,Myrf,Ndrg1,Gpr37,Gjb1,Pde8a,Cntn2,Gsn,Fgfr2,Scd2,Jam3,Abca2 | Plp1,Malat1,Ptgds,Scd2,Mbp,Fth1,Glul,Apod,Mobp,Mag,Cryab,Bin1,Ndrg1,Gpm6b,Mog,Ywhaq,Cd81,Gpr37,Tmsb4x,Hmgb1 |
4 | 101,101,255 | 0.08492 | 270686250 | Apoe,Clu,Slc1a2,Glul,Aldoc,Atp1a2,Slc1a3,Cst3,Gja1,Plpp3,Ckb,Ndrg2,Sparcl1,Gpr37l1,Gfap,Gstm1,Aqp4,Mt1,Mt3,Mfge8 | Gja1,Slc1a3,Aldoc,Apoe,Atp1a2,Clu,Gpr37l1,Aqp4,Plpp3,Slc1a2,Gfap,Mt1,Mfge8,Gstm1,Slc6a11,Ndrg2,Glul,Ednrb,Ntsr2,Agt | Apoe,Clu,Glul,Slc1a2,Cst3,Aldoc,Malat1,Ckb,Atp1a2,Sparcl1,Slc1a3,Mt3,Cpe,Ndrg2,Plpp3,Scd2,Tspan7,Gja1,Ntrk2,Dbi |
6 | 255,178,101 | 0.08061 | 256947664 | Pcp4,Itm2c,Calb2,Nnat,Pcsk1n,Rtn1,Map1b,Cacna1e,Thy1,Clstn1,Cbln2,Gabbr2,Tcf7l2,Bex1/2,Rit2,Kcnd2,Slc2a13,Apba1,Tac1,Nsg1 | Calb2,Cbln2,Tac1,Nnat,Oprm1,Tcf7l2,Pcp4,Apba1,Slc5a7,Rit2,Slc2a13,Sstr2,Rgs6,Itm2c,Cartpt,Synpr,Gabbr2,Pcsk1n,Gria4,Fgf1 | Itm2c,Rtn1,Pcp4,Malat1,Map1b,Pcsk1n,Cpe,Atp1b1,Aldoa,Tuba1a/b/c,Hsp90ab1,Thy1,Psap,Rtn3,Eif4a2,Clstn1,Gnao1,Meg3,Cacna1e,Slc25a4 |
7 | 178,255,101 | 0.04347 | 138561767 | Bsg,Slc2a1,Cldn5,Flt1,Itm2a,Rgs5,Pltp,Igfbp7,Serinc3,Esam,Vim,Slc7a5,Acta2,Igf1r,Id1,Fn1,Sptbn1,Srgn,Pecam1,St3gal6 | Flt1,Cldn5,Itm2a,Rgs5,Pltp,Acta2,Fn1,Pecam1,Esam,Slc2a1,Igfbp7,Myl9,Lsr,Emcn,Slc7a5,Kdr,Tagln,Srgn,St3gal6,Vim | Bsg,Malat1,Slc2a1,Tmsb4x,Serinc3,Cldn5,Sptbn1,H3f3b,Flt1,Myl6,Aplp2,Itm2a,Gnb1,Pltp,Rgs5,Cpe,Itm2b,Hmgb1,Calm1,Igfbp7 |
11 | 255,153,204 | 0.0429 | 136738469 | Ptprz1,Cspg5,Pdgfra,Olig1,Gpr17,Epn2,Tnr,Pllp,Vcan,Fyn,Serpine2,Cd9,Tuba1a/b/c,Olig2,Pcdh15,S100a16,S100b,Sulf2,Serinc5,Cntn1 | Vcan,Pdgfra,Gpr17,Pcdh15,Olig2,Tnr,S100b,Olig1,Ptprz1,Cspg5,Fyn,Epn2,Itpr2,Serpine2,Cd9,S100a16,Sox6,Megf11,Pllp,Sulf2 | Malat1,Tuba1a/b/c,Ptprz1,Cspg5,Olig1,Epn2,Pllp,Camk2a,Cd9,Serpine2,Calm1,Ncam1,Fyn,Bcan,Ckb,Tnr,Pdgfra,Cntn1,Hnrnpa2b1,Hmgb1 |
10 | 178,101,255 | 0.03812 | 121510243 | Ttr,Psap,Cab39l,Chchd10,Lamp2,Cox8a,Htr2c,Slc12a2,Ppp1r1b,Timp2,Ctsd,Mdh1,Gpi1,Bsg,Ndufa4,Itpr1,Ftl1,App,Sem1,Itgb8 | Ttr,Htr2c,Cab39l,Ppp1r1b,Lepr,Slc12a2,Col4a5,Foxj1,Lamp2,Msr1,Dlk1,Timp2,Ctnna1,Chchd10,Maob,Sil1,Itpr1,Tspo,Cd55,Dcn | Ttr,Psap,Malat1,Atp1b1,Cox8a,Mdh1,Bsg,Gpi1,Ctsd,Cab39l,Chchd10,App,Ndufa4,Clu,Ckb,Hspa8,Lamp2,Ptgds,Cst3,Slc12a2 |
9 | 101,178,255 | 0.03663 | 116755577 | Cst3,Camk2a,C1qc,Hexb,Ctsd,C1qa,C1qb,Ctss,Csf1r,Vtn,P2ry12,Sparc,Selplg,Cx3cr1,Hmgb1,Tyrobp,Itm2b,Rps9,Tmem119,Tgfbr1 | C1qc,Csf1r,Ctss,C1qb,C1qa,Hexb,Selplg,P2ry12,Tmem119,Cx3cr1,Tyrobp,Vtn,Trem2,Ptprc,Csf3r,Tgfbr1,Lyz1/2,Cd84,Higd1b,Epb41l2 | Cst3,Camk2a,Tmsb4x,Hmgb1,Ctsd,Itm2b,Malat1,Rps9,Sparc,Hexb,Glul,C1qc,C1qa,Fau,C1qb,Ctsb,Ctss,Csf1r,Vtn,Fth1 |
8 | 0,223,95 | 0.03466 | 110484904 | Mbp,Fth1,Mobp,Cpe,Bcas1,Hipk2,Kif5c,Pink1,Dync1li2,Gfap,Rims1,Cd6,Frs2,Lpar1,Ptpn11,Gpm6b,Map4k4,Ndrg1,S100b,Drd1 | Mbp,Mobp,Bcas1,Fth1,Hipk2,Pink1,Cd6,Cpe,Dync1li2,Kif5c,Gfap,Rims1,Frs2,Lpar1,Ptpn11,S100b,Map4k4,Drd1,Gpm6b,Ndrg1 | Mbp,Fth1,Mobp,Cpe,Malat1,Kif5c,Clu,Apoe,Glul,Plp1,Bcas1,Tmsb4x,Dync1li2,Hipk2,Pink1,Gfap,Hsp90ab1,Slc1a2,Gpm6b,Calm1 |
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.15824926.
These datasets can also be loaded directly using the following catalog YAML file:
https://zenodo.org/records/15824927/catalog.yaml