[awarkow]

Summary

This PR adresses #89 by adding sequencing-specific assay classes to the Common Access Model. They extend the existing generic Assay class so we can capture experiment metadata for whole genome, whole exome and RNA sequencing data.

New Assay Classes

The new classes form a hierarchy under the existing Assay class:
Assay
└── SequencingAssay
├── GenomicsAssay (WGS/WXS)
└── TranscriptomicsAssay (RNA-Seq)

Each subclass inherits all slots from its parent. Every sequencing assay therefore keeps the core Assay fields (assay_id, subject_id, sample_id, file_id, assay_type, etc.) and adds its own experiment metadata.

SequencingAssay

Sequencing Experiment Assay: a sequencing experiment performed on a Biospecimen. This is the shared base for all sequencing assays.

Slot	Range	Description
sequencing_center_id		The sequencing center that performed the experiment
library_name	string	Name of the sequencing library
experiment_date	date	Date the sequencing experiment was performed
experiment_strategy	EnumSequencingStrategy	Sequencing strategy used
is_paired_end	boolean	Whether reads are paired-end
platform	EnumSequencingPlatform	Sequencing platform used
instrument_model	EnumSequencingInstrument	Sequencing instrument model used
total_reads	integer	Total number of reads generated
is_adapter_trimmed	boolean	Whether adapter sequences were trimmed
adapter_sequencing	string	Adapter sequence used, if known (expected when is_adapter_trimmed is false)
quality_scale	EnumReadGroupQualityScale	Quality score scale used for the read group
reference_genome	EnumReferenceGenome	Reference genome build the data was aligned to or generated against
workflow_type	string	Type of workflow that produced the file (Kids First Dataservice: workflow_type)
workflow_tool	string	Tool used by the workflow (Kids First Dataservice: workflow_tool)
workflow_version	string	Version of the workflow tool
workflow_endpoint		Workflow endpoint
flow_cell	string	Flow cell identifier for the read group
lane_number	integer	Lane number for the read group

GenomicsAssay

WGS/WXS Assay: a whole genome or whole exome sequencing experiment performed on a Biospecimen. It extends SequencingAssay with alignment and coverage metrics.

Slot	Range	Description
max_insert_size	integer	Maximum insert size (bp)
mean_insert_size	float	Mean insert size (bp)
mean_depth	float	Mean sequencing depth of coverage
mean_read_length	integer	Mean read length (bp)
target_capture_kit	string	Target capture kit used, if any (relevant for WXS)

TranscriptomicsAssay

RNA Sequencing Assay: an RNA sequencing experiment performed on a Biospecimen. It extends SequencingAssay with library construction details.

Slot	Range	Description
library_strand	EnumLibraryStrand	Strandedness of the sequencing library
library_prep	EnumLibraryPrep	Library preparation method
library_selection	EnumLibrarySelection	Library selection method

New Enums

Caution

Work in progress. The new slots reference several enums. Only the 3 below are defined so far. Need to pair others with correct ontologies

EnumLibraryPrep

Library preparation methods for RNA sequencing. Values are drawn from OBI and constrained to subclasses of an OBI source node.

Value	Title	Description
OBI:0000848	polyA RNA extraction	An RNA extraction process, typically using poly dT oligomers, whose desired output material is polyA RNA.
OBI:0000855	total RNA depletion	An RNA extraction process in which total cellular and organelle RNA are extracted.

EnumAnalyteType

The type of analyte extracted from the biospecimen. Values are drawn from NCIt and constrained to subclasses of NCIt C706.

Value	Title	Description
C449	DNA	A long linear double-stranded polymer formed from nucleotides attached to a deoxyribose backbone and found in the nucleus of a cell; associated with the transmission of genetic information.
C812	RNA	Single-stranded long chain of nucleotides containing ribose.

EnumLibrarySelection

Library selection methods used to enrich or select nucleic acid for sequencing. Values are drawn from NCIt and constrained to subclasses of NCIt C15241.

Value	Title	Description
C163986	Hybrid Selection	A genomic library where the sample has been subjected to enrichment via nucleic acid hybridization.
C17003	PCR	A method for amplifying a DNA base sequence using multiple rounds of heat denaturation of the DNA and annealing of oligonucleotide primers complementary to flanking regions in the presence of a heat-stable polymerase.
C163987	Affinity Enrichment	A genomic library where the sample has been subjected to enrichment via binding to a target protein or oligonucleotide prior to molecular analysis.
C163988	Poly-T Enrichment	A genomic library where the sample has been subjected to enrichment via binding to immobilized poly-T oligonucleotides, which will preferentially isolate polyadenylated (poly-A) RNAs.
C163989	Random	A genomic library constructed to generate random DNA fragments by subjecting a genomic sample to either amplification using random primers or fragmentation via exposure to restriction enzymes or shear stress.
C163990	rRNA Depletion	A genomic library where the sample has been subjected to a ribosomal RNA depletion step before it is amplified.
C163991	miRNA Size Fractionation	A RNA library that has been fractionated by size exclusion methods to enrich for microRNAs.

Open Questions

• Is library_prep the same thing as library_construction? If so, should we rename or alias it?
• Should adapter_sequencing be conditionally required when is_adapter_trimmed is false? Can we enforce this with a LinkML rule?

[github-actions]

PR Preview Action v1.8.1
🚀 View preview at https://include-dcc.github.io/common-access-model/pr-preview/pr-113/
Built to branch gh-pages at 2026-11-05 18:30 UTC. Preview will be ready when the GitHub Pages deployment is complete.

[awarkow]

@torstees @yelenacox I used term-weaver to generate this enum (its a small subset but I wanted to test it out). The values are here in OLS and I was able to expand using weaver but for some reason the formatting is off. The values should be in the format ILLUMINA_1_3 but are appearing as ILLUMINA_1:3.

[RobertJCarroll]

Thanks for flagging this. We've been reviewing it. There's some "bad"/"probably unintended" behavior within OLS. We've been using the obo_id field as its' typically a nicely formatted CURIE, but for non-OBOs it is interpreting the _ as part of the OBO id, turning it (incorrectly) into a curie.
https://obofoundry.org/id-policy.html

FOUNDRY_OBO_URI ::= "http://purl.obolibrary.org/obo/" IDSPACE "_" LOCALID

There's not a consistent way to pull the data via API, but we have some ideas about solutions.

[awarkow]

🚧 Enums still in progress but need review

Paired to ontologies

• EnumAnalyteType -> OBI
• EnumLibraryPrep -> OBI
• EnumLibrarySelection -> NCI
• EnumLibraryStrand -> NCI
• EnumReadGroupQualityScale -> ValueSets

Still to pair

• EnumSequencingInstrument
• EnumSequencingCenter
• EnumSequencingPlatform
• EnumSequencingStrategy
• EnumReferenceGenome

[RobertJCarroll]

I started to look over the enums. I think it might be helpful if we documented some better conventions / easy paths to leveraging the tools as I can tell some things are unclear.

• The enumerated values should include the curie prefix, eg, ncit:C163986 not just C163986. This goes for the meaning as well.
• You only need reachable_from if you plan to use the integrated expansion. If you want to define it all "by hand", you can skip that field/section.
• You can "cheat" a little and get term weaver to do some heavy lifting. I see that a lot of these NCIt hierarchy sections aren't well specified / don't line up with your enums, but there are a few approaches to take that might speed up the process:

1. Pick a rough parent code in the reachable_from, run term-weaver, then manually prune the results and remove the reachable_from.
2. Pick a rough parent code in the reachable_from, add the minus list for things you don't want, then run term weaver. Minus is described on this page. You could also run term weaver with the rough parent, then pull the things you don't want into the minus list from that.
3. You can also just add all of your terms to source_nodes and make sure include_self: true. This will expand everything you included so you don't have to manually get the titles/descriptions. You can either clean up the reachable_from at that point or leave it in- just depends on if you "like" the query as it might add more terms if re-run in the future.

I'll work up some examples as I'm sure it would be helpful for any potential contributor.

[RobertJCarroll]

I took a look at EnumLibrarySelection. You can seed with:

---
id: https://includedcc.org/common-access-model/EnumLibrarySelection
name: EnumLibrarySelection
enums:
  EnumLibrarySelection:
    description: null
    reachable_from:
      source_ontology: bioregistry:ncit
      source_nodes:
        - NCIT:C15241
      include_self: false
      is_direct: false
      relationship_types:
        - rdfs:subClassOf

Which will grab everything but PCR. You could then add it in manually. If you did, I'd recommend making some note about the changes you made to help prevent future issues if someone wanted to regenerate. (edit: We could also discuss some potential additional features / options here?)

I'm a little suspicious of that PCR term though- how is that one distinct from the others?

[RobertJCarroll]

OLS seems to be struggling, but let me know if it'd be helpful to dig into these more. It should make writing the enums easier / cleaner!

[RobertJCarroll]

@awarkow We've been investigating the Valuesets "ontology"- turns out it's a collection of enumerations. While it is loaded into a few places, it's wonky in OLS. We're putting a bug report in to them about it, but we are also wondering how attached you are to using that list. Is there an underlying set of terms you all are using that we could leverage or are these the right ones? We are considering creating terms given the odd API behaviors.

[awarkow]

Summary of Added Enums

For review by Kids First BIX team: @yuankunzhu

Maps the sequencing-related constants from kf_lib_data_ingest/common/constants.py onto ontology-backed enums.

8/10 enums added in src/common_access_model/schema/enums/ are mapped. 2 still need discussion. Summarized here:

Enum	Source constant	Ontology
EnumSequencingPlatform	SEQUENCING.PLATFORM	NCIT
EnumSequencingInstrument	SEQUENCING.INSTRUMENT	OBI
EnumSequencingStrategy	SEQUENCING.STRATEGY	NCIT
EnumLibraryStrand	SEQUENCING.STRAND	NCIT
EnumLibrarySelection	SEQUENCING.LIBRARY.SELECTION	NCIT
EnumLibraryPrep	SEQUENCING.LIBRARY.PREP	OBI, NCIT
EnumAnalyteType	SEQUENCING.ANALYTE	SNOMED CT
EnumReferenceGenome	SEQUENCING.REFERENCE_GENOME	LOINC
---	---	---
EnumSequencingCenter	SEQUENCING.CENTER	CAMO?
EnumReadGroupQualityScale	READ_GROUP.QUALITY_SCALE	ValueSets

---

Mappings

EnumSequencingPlatform

Original value	Mapped to	Code
Complete Genomics	Complete Genomics Sequencing	NCIT:C146815
SOLiD	ABI SOLiD Sequencing	NCIT:C146811
Illumina	Illumina Sequencing	NCIT:C146817
LS454	454 Sequencing	NCIT:C146809
ONT	Nanopore Sequencing	NCIT:C146818
PacBio	SMRT Sequencing	NCIT:C146819
Ion Torrent	Ion Semiconductor Sequencing	NCIT:C125894

Added a reachable_from block (NCIT) listing these seven nodes with include_self: true.

EnumSequencingInstrument

Original value	Mapped to	Code
HiSeq X Ten	Illumina HiSeq X Ten	OBI:0002129
HiSeq X v2.5	not mapped (no matching OBI term found)	n/a

I also added the other Illumina HiSeq models found under OBI OBI:0003683, which are not in the original constants:
HiSeq 1000 (OBI:0002022), 1500 (OBI:0003386), 2000 (OBI:0002001), 2500 (OBI:0002002), 3000 (OBI:0002048), 4000 (OBI:0002049).
Added a reachable_from block rooted at OBI:0003683.

EnumSequencingStrategy

Original value	Mapped to	Code
WGS	Whole Genome Sequencing	NCIT:C101294
WXS	Whole Exome Sequencing	NCIT:C101295
RNA-Seq	RNA Sequence Analysis	NCIT:C192768
miRNA-Seq (MRNA)	mRNA Sequencing and MicroRNA Sequencing	NCIT:C129432, NCIT:C156057
scRNA-Seq	Single Cell RNA Sequencing	NCIT:C171152
Panel	Tumor Panel Sequencing	NCIT:C165600
	Disease Panel Gene Sequencing	NCIT:C173599
	Multi-gene Panel Sequencing	NCIT:C198412
Targeted Sequencing	Targeted Capture Sequencing	NCIT:C204828
	Targeted Genome Sequencing	NCIT:C158253
	Targeted Transcriptome Sequencing	NCIT:C158252
Methylation	Methyl-binding Domain Sequencing	NCIT:C204820
	Methylated DNA Immunoprecipitation Sequencing	NCIT:C204821
	Methylation-Sensitive Restriction Enzyme Sequencing	NCIT:C204822
Proteomics	Proteomics	NCIT:C20085
ONT WGS	Oxford Nanopore Sequencing	NCIT:C146818
CCS WGS, CLR WGS	Long Read Sequencing (one code covers both)	NCIT:C223895

No mapping yet (left as placeholder comments in the YAML): Linked-Read WGS (10x), snRNA-Seq, CCS RNA-Seq, CLR RNA-Seq.

EnumLibraryStrand

Original value	Mapped to	Code
Unstranded	Unstranded Library Sequencing	NCIT:C150427
First Stranded	First-strand Library Sequencing	NCIT:C150428
Second Stranded	Second-strand Library Sequencing	NCIT:C150426
Stranded	not mapped (generic term, no matching code)	n/a

EnumLibrarySelection

Original value	Mapped to	Code
Hybrid Selection	Hybrid Selected Genomic Library	NCIT:C163986
Affinity Enrichment	Affinity Enriched Genomic Library	NCIT:C163987
Poly-T Enrichment	Poly-A Enriched Genomic Library	NCIT:C163988
Random	Randomly Selected Genomic Library	NCIT:C163989
rRNA Depletion	rRNA Depleted Genomic Library	NCIT:C163990
miRNA Size Fractionation	miRNA Size Fractionated Genomic Library	NCIT:C163991
PCR	Polymerase Chain Reaction	NCIT:C17003

EnumLibraryPrep

Original value	Mapped to	Code
polyA	polyA RNA extraction	OBI:0000848
	Poly-A Enriched Genomic Library	NCIT:C163988
totalRNAseq	total RNA depletion	OBI:0000855
	rRNA Depleted Genomic Library	NCIT:C163990

EnumAnalyteType

Original value	Mapped to	Code
DNA	DNA	SNOMEDCT:24851008
RNA	RNA	SNOMEDCT:27888000
Virtual	not mapped	n/a

EnumReferenceGenome

Original value	Code
GRCh37	loinc:LA14029-5
GRCh38	loinc:LA26806-2
hg19	loinc:LA14025-3
hs37d5	commented out (no code found)

---

Questions for reviewers

1. MRNA = "miRNA-Seq" upstream. The constant name says mRNA but the value says miRNA, so I mapped it to both NCIT:C129432 (mRNA) and NCIT:C156057 (microRNA). Should one of these be dropped?
2. One-to-many strategy mappings. Methylation, Targeted and Panel each map to several NCIT terms. Is only one of these correct or should all be included as potential values?
3. Long-read strategies. CCS WGS and CLR WGS collapse to one Long Read Sequencing code, so the CCS/CLR distinction is lost. The RNA variants are unmapped since the description mentions sequencing of DNA. Additional values may be needed in CAMO.
4. Unmapped values. HiSeq X v2.5, Virtual, hs37d5, Linked-Read WGS, snRNA-Seq, CCS RNA and CLR RNA have no code yet. Suggestions welcome.
5. EnumLibraryPrep overlaps with EnumLibrarySelection. NCIT:C163988 and NCIT:C163990 appear in both enums. Is that correct?
6. NCIT:C146818 in two enums. It is the platform "Nanopore Sequencing" in EnumSequencingPlatform and "Oxford Nanopore Sequencing" in EnumSequencingStrategy, so the same code is used for ONT WGS. Is this okay?

[awarkow]

@RobertJCarroll That list (ValueSets/SequenceQualityEncoding) contains the exact 5 values we need to add for this enum. I thought it had potential to use since it does match exactly with our preferred terms, but I'm okay with adding to CAMO if the formatting in OLS is messing things up

[RobertJCarroll]

This is where we are talking with the OLS team as an fyi: EBISPOT/ols4#1484

[yuankunzhu]

@awarkow for the questions you have

• MRNA = "miRNA-Seq" upstream. The constant name says mRNA but the value says miRNA, so I mapped it to both NCIT:C129432 (mRNA) and NCIT:C156057 (microRNA). Should one of these be dropped?

  • drop the mRNA code, the value in the data is "miRNA-Seq" which means the microRNA, so we should just map it to NCIT:C156057

• One-to-many strategy mappings. Methylation, Targeted and Panel each map to several NCIT terms. Is only one of these correct or should all be included as potential values?

  • Panel: we should use NCIT:C198412 Multi-gene Panel Sequencing, to be more general, it's the parent of of both Tumor Panel and Disease Panel
  • Targeted Sequencing: we typically don't have Targeted Transcriptome Sequencing, and the "Targeted Genome Sequencing" is the parent of "Targeted Capture Sequencing", so I suggest we go with the parent term
    "Targeted Genome Sequencing | NCIT:C158253"
  • Methylation: for now, we have only had array based methylation data, so all three listed terms are wrong, we should use NCIT:C165222 DNA Methylation Array

• Long-read strategies. CCS WGS and CLR WGS collapse to one Long Read Sequencing code, so the CCS/CLR distinction is lost. The RNA variants are unmapped since the description mentions sequencing of DNA. Additional values may be needed in CAMO.

  • CCS and CLR are the different sequencing mode for PacBio as pose to a "Sequencing Strategy"
  • also noticed we listed "ONT WGS" as "Oxford Nanopore Sequencing" under "Sequencing Strategy", which could also be miss-leading, because Nanopore can be used for WGS, RNA-seq (including direct RNA), targeted panels and so on
  • I will suggeset the following:
    • add "CCS" and "CLS" as a enum values for "sequencing_mode" in CAMO and then apply the following
    • ONT WGS: strategy C101294 (WGS) + platform C146818 (Nanopore)
    • CCS WGS: strategy C101294 (WGS) + platform C146818 (PacBio) + sequencing_mode (CCS)
    • CLR WGS: strategy C101294 (WGS) + platform C146819 (PacBio) + sequencing_mode (CLS)
    • i think this should also address your last question on "NCIT:C146818 in two enums"
      Happy to discuss more about this.

• EnumLibraryPrep overlaps with EnumLibrarySelection. NCIT:C163988 and NCIT:C163990 appear in both enums. Is that correct?

  • keep LibraryPrep OBI-only (polyA as OBI:0000848, and totalRNAseq as OBI:0000855) and keep the NCIt library terms in LibrarySelection. C124261 should come out of LibraryPrep, because it’s a strategy term.