CHANGELOG.md

Changelog

All notable changes to scpca-nf are documented here.

The format is based on Keep a Changelog, and this project adheres to Semantic Versioning.

Unreleased

Added

• Added processes to perform formatting checks for SCE and AnnData objects generated by the workflow
• New Docker containers for scvi-tools with experimental GPU support(currently unused)

Changed

• Added a minimum cell threshold for running SCimilarity
• H5AD conversion for CellAssign and SCimilarity is now done in a separate process to allow for separation of R and Python processes.

Fixed

Deprecated

Removed

v0.10.5

Added

• Accommodate other columns besides cell_type_annotation that may be present in the submitter cell types file. All columns present in the input file are now included in the cell metadata of all objects.

v0.10.4

Added

• Included support for running gene expression libraries from 10x Multiome (GEX + ATAC) v1 kit

v0.10.3

Added

• New citation for the published manuscript
• Added guidelines for development with Claude Code in CLAUDE.md

Changed

• Skip scDblFinder for libraries with a low number of cells
• Cell typing is only performed if the organism listed is Homo sapiens
• Update handling of FASTQ file names for Space Ranger

Fixed

• mapping tool is now cellranger multi instead of cellranger-multi for GEM-X 10x Flex libraries

0.10.2 — 2026-05-05

Fixed

• Fix type error for numerical clusters when making markers

0.10.1 — 2026-05-05

Changed

• Updated display of project list and marker gene lists in Cell Browser

Fixed

• Fixed a bug that caused extra reruns of inferCNV
• Grammar update in report

0.10.0 — 2026-04-07

• Updated spatial transcriptomics processing to use Space Ranger v4; supported Visium technologies now include:
  • First-generation Visium
  • Visium CytAssist
  • Visium HD and Visium HD 3'
• Added support for additional Visium Flex Gene Expression probe set versions
• Updated Cell Browser output to display normalized expression values instead of raw counts
• Addressed a small error in reference building
• Updated to scpcaTools v0.4.5, containing minor R version and infrastructure updates
• Infrastructure updates:
  • Cell Ranger and Space Ranger Docker containers used by scpca-nf are now hosted in this repository
  • Added pull-through cache support for Docker containers
• Initiated scpca-nf object formatting checks and updated internal metrics report

0.9.3 — 2026-02-04

• Refactored counting reference cells for InferCNV; the infercnv_success boolean in object metadata has been replaced with the infercnv_status string
• Minor QC report updates to adjust dot plot sizing based on the number of validation markers for consensus cell types
• Updated the consensus cell type reference to use OpenScPCA-analysis v0.2.4, fixing a bug where consensus cell types were incorrectly unassigned if 2/3 methods agreed but the third was unable to assign a cell type
• Validation palette now contains colors for all possible validation groups as of OpenScPCA-analysis v0.2.4

0.9.2 — 2025-12-04

• Minor bug fix in build-index.nf
• Ensure Cell Browser always rebuilds in build-cellbrowser.nf
• Updated Cavatica configuration files and documentation

0.9.1 — 2025-11-17

• CNV inference is now skipped for samples with a diagnosis of Non-cancerous, even when --perform_cnv_inference is used
• Added additional checks to improve handling of unknown diagnoses and QC report rendering when running with CNV inference
• All technology values are now converted to lowercase when reading from existing checkpoint files
• Added an explicit check that technology has not changed when skipping the mapping steps for a library

0.9.0 — 2025-11-05

• Added support for additional 10x Genomics kits:
  • Flex Gene Expression (single and multiplexed)
  • Chromium GEM-X single-cell 5' v3
  • Chromium GEM-X single-cell 3' v4
• Doublet detection now performed using scDblFinder; results stored in filtered and processed objects
• --perform_celltyping now additionally assigns cell types using SCimilarity alongside SingleR and CellAssign
  • Consensus cell types determined by agreement of ≥2 of 3 methods in an ontology-aware manner using the updated consensus reference in OpenScPCA-analysis v0.2.3
• Cell types annotated as part of the OpenScPCA Project can now be provided as input; annotations are added to cell metadata of all objects and summarized in QC and cell type reports
• Added --perform_cnv_inference flag for CNV inference using a reference of non-malignant cells determined by consensus cell types
• Added new build-cellbrowser.nf workflow to build a UCSC Cell Browser instance from main workflow output
• Added Cavatica instructions

0.8.8 — 2024-07-15

• Updated to scpcaTools v0.4.3, which fixes a bug where cell_id in merged objects now correctly includes both library ID and cell barcode
• Bug fix for #867

0.8.7 — 2024-04-14

• Bulk workflow now outputs both raw counts and TPM
• Defined resource limits for individual processes
• Updated manifest and schema formatting
• Added two new profiles: staging and production
• Output a metrics JSON file
• Processed and merged objects now contain a consensus cell type assignment
• Updated to scpcaTools v0.4.2, which saves assays in merged objects as sparse matrices rather than DelayedArray

0.8.6 — 2024-01-28

• merge.nf now requires a minimum of 3 cells for merging
• Updated GHA artifacts to v4
• Added support for samples categorized as both a patient-derived xenograft and a cell line

0.8.5 — 2023-11-04

• Updated to scpcaTools v0.4.1:
  • Raw single-cell counts are now correctly rounded
  • Security updates via upgraded lightning in dependent Python packages
• Converted deprecated addParams() to take for future Nextflow compatibility
• Fixed handling of the 2-cell case to allow PCA creation

0.8.4 — 2023-08-09

• Only count single-cell libraries toward the max merged limit
• Account for different HVG name in merged object

0.8.3 — 2023-08-05

• Updated AnnData objects:
  • PCA and UMAP embeddings now stored as X_pca and X_umap (previously X_PCA and X_UMAP)
  • .var now contains a highly_variable column
  • .uns now contains a pca object with parameters and variance weights
• Added max_merge_libraries parameter for merge.nf to limit the number of libraries merged into a single object

0.8.2 — 2023-07-16

• Updated to use slimmer scpcaTools Docker containers (v0.4.0)
• Bulk files (bulk_quant.tsv and bulk_metadata.tsv) are now saved to a bulk directory
• QC and cell type reports now include a table summarizing sample metadata (age, sex, diagnosis, etc.)

0.8.1 — 2023-04-29

• Output HDF5 files now use the .h5ad extension instead of .hdf5
• Bug fixes:
  • Skip CellAssign if processed object has only 1 cell
  • No cell type reports generated if processed objects have only 1 cell
  • Account for no genetic demultiplexing when estimating sample cell counts for metadata.json
  • Skip downstream processing if no cells remain after filtering empty droplets

0.8.0 — 2023-03-15

• Added merge.nf workflow for creating merged SingleCellExperiment and AnnData objects from a list of projects, including a summary report
• Minor aesthetic fixes to cell type report
• Use bz2 compression for SingleCellExperiment objects to reduce file sizes
• Added nextflow_schema.json to the repository

0.7.3 — 2023-03-11

• Skip clustering and cell type assignment for processed objects with 0 cells
• Use new max_mem label for CellAssign classification, using params.max_memory after first failure

0.7.2 — 2023-02-05

• Skip CellAssign if too few cells are present
• Bug fixes for running multiplexed samples through cell typing
• Skip cell typing for cell line samples
• Adjustments to UMAPs in the cell type report
• Account for missing input files

0.7.1 — 2023-01-03

• Updated memory and CPU requests for CellAssign
• Set max retries after AWS instance failure to 2
• Use On Demand instances on 3rd retry
• Added faceted UMAPs to the cell type report

0.7.0 — 2022-11-28

• Added cell type annotation to the main workflow:
  • Annotation with SingleR or CellAssign
  • New output includes processed SingleCellExperiment and AnnData objects with cell type annotations and a cell type summary report
  • Documentation for adding cell type annotations
• Cell type annotation is off by default; use --perform_celltyping to enable
• By default, existing cell typing results are reused; use --repeat_celltyping to rerun
• Added pre-commit to the repository

0.6.3 — 2022-11-17

• Ensure all RDS files are compressed
• Support running multiple projects at once via the --project parameter

0.6.2 — 2022-11-08

• Fixed a bug to ensure at least one gene or feature is present in the SingleCellExperiment object prior to AnnData conversion
• Temporarily removed cell type section from the main QC report

0.6.1 — 2022-10-30

• Fixed a bug when handling NA values in library metadata

0.6.0 — 2022-10-12

• Added cluster assignments to processed SingleCellExperiment objects
• Incorporated sample metadata into SingleCellExperiment objects
• Export all SingleCellExperiment objects as AnnData objects
• Added a workflow for generating references for cell type annotation
• Added CellAssign for cell type annotation
• Initial cell type report additions
• Allowed addition of submitter-provided cell types
• Added style guide
• Handle failures in modeling gene variance and PCA

0.5.4 — 2022-09-14

• Added error handling for modeling gene variance, calculating PCA, and calculating UMAP

0.5.3 — 2022-07-19

• Fixed intermittent ADT filtering bug
• Improved memory usage for QC report

0.5.2 — 2022-07-11

• Added post-processing (filtering and normalization) for ADT data from CITE-seq experiments
• Added support for additional organisms as references
• QC report updates
• Workflow development updates: improved behavior in limited-resource environments; added more stub block testing
• Experimental features: sample integration; cell type annotation

0.5.1 — 2022-05-31

• Made genetic demultiplexing optional

0.5.0 — 2022-05-12

• Added support for processing libraries from multiple organisms
• Available mapping references now include Human and Mouse Ensembl v104
• Option to skip feature mapping if already completed
• Publish checkpoints for feature mapping

0.4.2 — 2022-05-05

• Updated URL
• Ensured feature barcode files provided to salmon index contain the correct columns

0.4.1 — 2022-01-11

• Catch errors in UMAP calculation in post_process_sce.R
• Updated to scpcaTools v0.2.1

0.4.0 — 2021-12-07

• Added filtering, normalization, and dimensionality reduction of SingleCellExperiment objects
• Metadata tracking at checkpoints
• Added script for downloading local reference files and images
• Added instructions for downloading local reference files
• Updated to scpcaTools v0.2.0, including updates to reading alevin-fry data with fishpond::loadFry()
• Reorganized documentation

0.3.4 — 2021-09-28

• Fixed potentially missing (unneeded) multiplex pool file
• Made scripts executable
• Added AWS anonymous access to example config for easier running on external systems

0.3.3 — 2021-06-13

• Updated to scpcaTools v0.1.8, including special handling of libraries with one sample and cell hashing in the QC report
• Skip demultiplexing for cell-hashed libraries containing only one sample

0.3.2 — 2021-06-06

• Updated to scpcaTools v0.1.7, which adds a dedicated section in the QC report for multiplexed libraries

0.3.1 — 2021-05-26

• Updated Alevin-fry to 0.5.0 and Salmon to 0.8.0
• Improved error handling for miQC failures
• Changed default queue to use larger disk after first job failure
• Added a new profile using auto-scaling EBS

0.3.0 — 2021-05-04

• Initial support for multiplexed libraries:
  • Added genetic demultiplexing workflow using cellsnp/vireo
  • Report demultiplexing results based on cellhash data
• Updated user instructions and example metadata

0.2.7 — 2021-04-12

• Spatial transcriptomics workflow output now includes the <library_id>_metadata.json file nested inside the <library_id>_spatial folder

0.2.6 — 2021-04-07

• Fixed a bug in the spaceranger workflow to ensure the index name is present in the metadata file even when the spaceranger process is skipped

0.2.5 — 2021-03-31

• Added option to skip spaceranger quantification for the spatial workflow using --repeat_mapping
• Reorganized spaceranger output file structure
• Added a ccdl profile

0.2.4 — 2021-03-17

• Split process settings from AWS profile
• Added minimal parameter checks and manifest
• Added template config for external users
• Added instructions for external users to README

0.2.3 — 2021-02-18

• Increased maximum retry attempts from 2 to 3

0.2.2 — 2021-02-17

• Increased memory request for fastp in the bulk RNA-seq workflow to accommodate larger libraries
• Memory request now doubles on every retry attempt

0.2.1 — 2021-02-15

• Split CPU and memory handling with labels
• Updated error handling for cell filtering using scpcaTools v0.1.4
• Added filtering method to metadata.json

0.2.0 — 2021-02-07

• Mapping steps are now skipped by default for bulk and single-cell RNA-seq data; use --repeat_mapping to force remapping
• Removed implicit dependence on S3 for file locations
• Moved index and barcode defaults to public reference bucket
• Added bulk RNA-seq metadata file per project
• Updated to scpcaTools v0.1.3, which rounds expression counts for single-cell data

0.1.3 — 2021-01-14

• Added bulk RNA-seq processing with Salmon
• Added spatial expression analysis with Space Ranger
• Memory and disk space updates for large samples
• Reorganized configuration files and parameters
• Updated to scpcaTools v0.1.2, changing filtering to use emptyDropsCellRanger

0.1.2 — 2020-10-18

• Changed default alevin-fry resolution to cr-like-em
• Reduced default memory usage

0.1.1 — 2020-10-08

• More robust handling of miQC model fit failures via scpcaTools v0.1.1

0.1.0 — 2020-10-07

Initial release of the scpca-nf workflow.