AlanSimmons
commented
8 months ago Change log - HUBMAP
EFO Alignment (tangential to soft assay)
The following was required to allow for multiple hierarchies of dataset.
Nodes in range HUBMAP:C003000 to HUBMAP:C003032 had been created to organize datasets hierarchically at the request of a group led by Katy Börner, described as an attempt to align assays with EFO. The stated use case was a special Sankey chart. This hierarchy has yet to be used.
I created a new parent node for this hierarchy (HUBMAP:C002099), named EFO Alignment Dataset Hierarchy. I also related the first-level child nodes (e.g., HUBMAP:C003000) to HUBMAP:C002099. All existing "Dataset" nodes (children of HUBMAP:C000004) are also in the EFO Alignment Dataset Hierarchy.
Soft Assay Dataset Category, Soft Assay Dataset Type
- HUBMAP:C003040 = Soft Assay Dataset Category, isa HUBMAP:C000004 (HuBMAP Dataset). Corresponds to Column A of the Pipeline Decision Rules document. I was informed that this is an arbitrary grouping; however, if I am asked to model something in UBKG that has a grouping column, I'm going to create a node hierarchy for that column in UBKG.
- HUBMAP:C003041 = Soft Assay Dataset Type, isa HUBMAP:C000004 (HuBMAP Dataset). Corresponds to Column D.
- HUBMAP:C003042 through HUBMAP:C003051 - isa HUBMAP:C003040 (Soft Assay Dataset Category)
- HUBMAP:C003052 through X
- isa HUBMAP:C003041 (Soft Assay Dataset Type)
- mapped via isa to appropriate category. For example, HUBMAP:C003052 (RNASeq) isa HUBMAP:C003043 (RNA Sequencing)
Existing (v1) Dataset data_type:Soft Assay Data Type mappings
In v1, each dataset is associated with a data_type (also called assay_type). Following are associations between existing datasets and the new Soft Assay Dataset Type, based on the v1 values of data_type.
The existing set of v1 values of data_type can be obtained with the API endpoint: https://ontology.api.hubmapconsortium.org/valueset?parent_sab=HUBMAP&parent_code=C004001&child_sabs=HUBMAP
Datasets in bold were not mentioned in the Pipeline Decision Rules spreadsheet.
| v1 Display Name in Portal | data_type | alt-names | Primary/Processed | Soft Assay Dataset Type | |
|---|---|---|---|---|---|
| Visium | Visium | none | Primary | Visium | |
| sciRNA-seq [Salmon] | salmon_rnaseq_sciseq | none | Processed | RNASeq | |
| Bulk RNA-seq [Salmon] | salmon_rnaseq_bulk | none | Processed | RNASeq | |
| scRNA-seq (10x Genomics) [Salmon] | salmon_rnaseq_10x | none | Processed | RNASeq | |
| snRNA-seq [Salmon] Dataset | salmon_sn_rnaseq_10x | salmon_rnaseq_10x_sn | Processed | RNASeq | |
| sciATAC-seq [SnapATAC] | sc_atac_seq_sci | none | Primary | ATACSeq | |
| Bulk ATAC-seq [BWA + MACS2] | bulk_atacseq | none | Processed | ATACSeq | |
| CODEX [Cytokit + SPRM] | codex_cytokit | none | Processed | CODEX | |
| CODEX [Cytokit + SPRM] | codex_cytokit_v1 | none | Processed | CODEX | |
| 10x Multiome | 10X Multiome | Primary | none | 10x Multiome | |
| Multiplexed IF Microscopy | MxIF | none | Primary | CyCIF | |
| Cell DIVE | cell-dive | cell DIVE, Cell DIVE | Primary | Cell DIVE | |
| CellDIVE [DeepCell + SPRM] | celldive_deepcell | none | Processed | Cell DIVE | |
| PAS Stained Microscopy | PAS | PAS microscopy | Primary | Histology | |
| MALDI IMS | MALDI-IMS | MALDI-IMS-neg, MALDI-IMS-pos | Primary | MALDI | |
| SIMS-IMS | SIMS-IMS | SIMS | Primary | SIMS | |
| DESI | DESI | DESI-IMS, DESI IMS | Primary | DESI | |
| NanoDESI IMS | NanoDESI | none | Primary | DESI | |
| Multiplex Ion Beam Imaging | MIBI | none | Multiplex Ion Beam Imaging, mibi | Primary | MIBI |
| Multiplex Ion Beam Imaging [DeepCell + SPRM] | mibi_deepcell | none | Processed | MIBI | |
| Imaging Mass Cytometry (2D) | IMC | 2D-IMC, Imaging Mass Cytometry | Primary | 2D Imaging Mass Cytometry | |
| LC-MS | LC-MS | none | Primary | LC-MS | |
| Label-free LC-MS | lc-ms_label-free | none | Primary | LC-MS | |
| Labeled LC-MS | lc-ms_labeled | none | Primary | LC-MS | |
| Label-free LC-MS/MS | lc-ms-ms_label-free | none | Primary | LC-MS | |
| Untargeted LC-MS | LC-MS-untargeted | none | Primary | LC-MS | |
| Bulk RNA-seq | bulk-RNA | none | Primary | RNASeq | |
| snATACseq (SNARE-seq2) | SNARE-ATACseq2 | SNAREseq, SNARE-seq2, SNARE2-ATACseq | Primary | ATACSeq | |
| snRNAseq (SNARE-seq2) | SNARE-RNAseq2 | SNARE2-RNAseq | Primary | RNASeq | |
| snATAC-seq (SNARE-seq2) [Lab Processed] | sc_atac_seq_snare_lab | none | Primary | ATACSeq | |
| snRNA-seq (SNARE-seq2) [Lab Processed] | sc_rna_seq_snare_lab | none | Primary | RNASeq | |
| snRNA-seq (SNARE-seq2) [Salmon] | salmon_rnaseq_snareseq | none | Processed | RNASeq | |
| snATAC-seq (SNARE-seq2) [SnapATAC] | sc_atac_seq_snare | none | Processed | ATACSeq | |
| scRNA-seq (10x Genomics v2) | scRNAseq-10xGenomics-v2 | none | Primary | RNASeq | |
| scRNA-seq (10x Genomics v3) | scRNAseq-10xGenomics-v3 | none | Primary | RNASeq | |
| sciATAC-seq | sciATACseq | none | Primary | ATACSeq | |
| sciRNA-seq | sciRNAseq | none | Primary | RNASeq | |
| snATAC-seq | snATACseq | none | Primary | ATACSeq | |
| snATAC-seq [SnapATAC] | sn_atac_seq | sn_atac_seq_multiome_10x | Primary | ATACSeq | |
| snRNA-seq (10x Genomics v2) | snRNAseq-10xGenomics-v2 | snRNAseq-v2 | Primary | RNASeq | |
| snRNA-seq (10x Genomics v3) | snRNAseq-10xGenomics-v3 | snRNAseq, snRNAseq-v3 | Primary | RNASeq | |
| Slide-seq | Slide-seq | none | Primary | RNASeq | |
| Slide-seq [Salmon] | salmon_rnaseq_slideseq | none | Processed | RNASeq | |
| Targeted Shotgun / Flow-injection LC-MS | Targeted-Shotgun-LC-MS | none | Primary | LC-MS | |
| TMT LC-MS | TMT-LC-MS | none | Primary | LC-MS | |
| LC-MS Bottom Up | LC-MS_bottom_up | LC-MS Bottom-Up | Primary | LC-MS | |
| LC-MS Top Down | LC-MS_top_down | LC-MS Top-Down | Primary | LC-MS | |
| Autofluorescence Microscopy | AF | none | Primary | Auto-fluorescence | |
| Lightsheet Microscopy | Lightsheet | none | Primary | Lightsheet | |
| Bulk ATAC-seq | ATACseq-bulk | bulkATACseq | Primary | ATACSeq |
Unmapped existing v1 data_type
| v1 Display Name in Portal | data_type | alt-names | Primary/Processed |
|---|---|---|---|
| DART-FISH | DART-FISH | none | Primary |
| image_pyramid | image_pyramid | none | n/a |
| Imaging Mass Cytometry (3D) | IMC3D | 3D-IMC, 3D Imaging Mass Cytometry | Primary |
| NanoPOTS | NanoPOTS | none | Primary |
| seqFISH | seqFISH | none | Primary |
| seqFISH [Lab Processed] | seqFish_lab_processed | none | Primary |
| Whole Genome Sequencing | WGS | none | Primary |
| MS | MS | none | Primary |
| MS Bottom Up | MS_bottom_up | MS Bottom-Up | Primary |
| MS_top_down | MS_top_down | MS Top-Down | Primary |
| Publication | publication | none | n/a |
| Publication ancillary | publication_ancillary | none | n/a |
| GeoMX | GeoMX | none | Primary |
| Kaggle-1 Glomerulus Segmentation Dataset | pas_ftu_segmentation | none | Processed |
New pipelines from Column B
In v1, these would have been considered values of data_type for derived datasets. A new Dataset node in UBKG would have been created, mapped to a Dataset Data Type node with value=the new pipeline.
The plan is not to add information on these pipelines to UBKG.
| Pipeline | Dataset Type |
|---|---|
| salmon_rnaseq_10x_v2 | RNASeq |
| salmon_rnaseq_10x_v2_sn | RNASeq |
| sc_atac_seq_sn | ATACSeq |
New Datasets from Column C
These have not yet been defined in UBKG.
- Visium no probes
- Visium with probes
- Visium (with probes)
- Visium 🦄with probes🦄
- 10x v2 scRNAseq
- 10x v2 snRNA
- PhenoCycler
- MERFISH
- H&E
- nanoSPLITS
- Confocal
- Thick section Multiphoton MxIF
- Second Harmonic Generation (SHG)
- Enhanced Stimulated Raman Spectroscopy (SRS)
- Molecular Cartography
Overloaded concepts
Node terms in the SimpleKnowledge spreadsheet used to build HUBMAP have to be unique. A consequence of this is that some nodes are of multiple type. (Note: "Dataset Data Type" is a kind of node that corresponds to the v1 _datatype property)
- HUBMAP:C014103 (Visium)
- from v1: isa Dataset Display Name,Dataset Data Type
- now: isa Soft Assay Dataset Type,Spatial Transcriptomics
- HUBMAP:C014004 (10x Multiome)
- from v1: isa Display Name
- now: isa Soft Assay Dataset Type,Soft Assay Dataset Category,10x Multiome
- Note: the concept has a isa relationship with itself, because the string "10x Multiome" is assigned to both a category and data type.
- HUBMAP:C006503 (CODEX)
- from v1: isa Dataset Data Type,Dataset Display Name
- now: isa Soft Assay Dataset Type,MxFBE
- HUBMAP:C006303 (Cell DIVE)
- from v1: isa Dataset Display Name
- now: isa Soft Assay Dataset Type,MxFBE
- HUBMAP:C003047 (Histology)
- isa Soft Assay Dataset Category,Soft Assay Dataset Type,Histology
- Note: the concept has a isa relationship with itself, because the string "Histology" is assigned to both a category and a data type.
- HUBMAP:C007804 (MIBI)
- v1: isa Dataset Data Type
- now: isa Soft Assay Dataset Type,MxNF
- HUBMAP:C011902 (LC-MS)
- v1: isa Dataset Display Name,Dataset Data Type
- now: isa Soft Assay Dataset Type
- HUBMAP:C003051 (Molecular Cartography) isa Soft Assay Dataset Category,Soft Assay Dataset Type,Molecular Cartography
- HUBMAP:C007604 (Lightsheet)
- v1: isa Dataset Data Type
- now: Soft Assay Dataset Type,Single-cycle Flourescence Microscopy
- HUBMAP:C012502 (DESI)
- v1: isa Dataset Data Type,Dataset Display Name
- now: isa Soft Assay Dataset Type
Other changes
- HUBMAP:C004020 term "maldi" changed to "maldi_vitessce_hint" so that the string "MALDI" could be used as a dataset type.
- HUBMAP:C002037 (cells) - synonym "histology" removed so that the string "Histology" can be used as a dataset type.
Statement of problem
Datasets are being reorganized to support dynamic pipeline assignment--aka "soft assay type".
Reference
In the old infrastructure ("hard assay types", I guess), a static data_type/assay_type was created for every new combination of primary dataset and pipeline. For example, salmon_rnaseq_10x_v2 is a data_type for the result of a pipeline processing of primary datasets with data_type scRNAseq-genomics-v2.
In the new infrastructure,
Proposed solution
Optional: