Bioconductor 3.22 Released
59 new software packages, updates for genomics, single-cell transcriptomics, spatial omics, QC tooling, core I/O, GFF/CIGAR handling, much more.
The Bioconductor 3.22 release is now available. It includes 2,361 software packages, 435 experiment data packages, 926 annotation packages, 29 workflows, and 6 books. This cycle adds 59 new software packages, 6 new experiment data packages, 2 new annotation packages, and 1 new book, along with many updates across the existing ecosystem.
A few themes stand out for people working in genomics, single-cell transcriptomics, and spatial omics. Interoperability for single-cell data continues to improve, both across file formats and R data structures. Spatial analysis gains more upstream pipelines and QC tooling, including options tuned for imaging-based platforms. Core I/O also gets attention, with focused utilities for GFF and CIGAR handling, and a dedicated package for the Seqinfo class that many Bioconductor objects rely on.
For single-cell data workflows, anndataR brings native R support for AnnData, including backed H5AD and Zarr, and convenient conversion to SingleCellExperiment and Seurat objects. That is useful if you keep analysis in R but receive data produced by Python pipelines. anglemania focuses on selecting genes for integration across batches by exploiting gene-gene correlations that are stable across datasets, a practical alternative to relying only on highly variable genes. For modeling dynamic expression along trajectories, scLANE provides spline-based negative binomial GLMs, GEEs, and GLMMs over pseudotime or latent time, with downstream clustering and enrichment utilities.
Spatial omics gets several helpful additions. stPipe provides an upstream pipeline for sequencing-based spatial transcriptomics, starting at FASTQ trimming and running through UMI deduplication, count matrix generation, QC, and visualization with a configuration-first design that favors reproducibility. Imaging-based platforms benefit from SpaceTrooper, which uses GLM models to compute QC metrics, detect outliers, and visualize polygons for transcriptomic or proteomic images. If you need to evaluate clustering results while accounting for spatial structure, spARI implements a spatially aware adjusted Rand index that is tailored for spatial transcriptomics data, which can be more informative than standard cluster agreement scores.
On the regulatory and 3D genomics side, linkSet standardizes representation and analysis of genomic interactions with an emphasis on promoter-enhancer relationships. It supports conversion from common interaction formats, annotation of promoters and enhancers, distance-based analyses, and tailored visualizations. HiCaptuRe provides an end-to-end toolkit for Capture Hi-C data, including import, annotation, manipulation, and export at restriction fragment resolution with features for comparing samples and conditions. HiCPotts takes a Bayesian approach to Hi-C by using a Potts model that captures spatial dependencies across interaction loci while modeling key covariates such as genomic distance, GC content, transposable elements, and accessibility. The result is a principled segmentation of interactions into noise, genuine signal, and possible false signal, with posterior summaries that make the modeling choices transparent.
Several new packages strengthen core infrastructure. Bioc.gff offers lightweight parsing and writing for GFF and GTF, which is handy when you want focused I/O without pulling in a larger stack. cigarillo exposes low-level utilities for parsing and projecting CIGAR strings, which many users will touch indirectly through higher-level packages but which can be valuable for custom alignment workflows. Seqinfo now ships as its own package, reflecting the central role of sequence metadata across GRanges, VCF, and many other Bioconductor classes.
The Seqinfo class was moved from the GenomeInfoDb package to the new Seqinfo package. IMPORTANT WARNING: This means that: Seqinfo-holding S4 objects that were serialized before this change (i.e. prior to Bioconductor 3.22) are no longer valid objects in Bioconductor >= 3.22. […] This also means that Seqinfo-holding S4 objects serialized with Bioconductor >= 3.22 are no longer valid objects in Bioconductor < 3.22.
The GenomeInfoDb package now depends on new Seqinfo package, and the following functions have moved to the new Seqinfo package:
constructor functions Seqinfo() and GenomeDescription()
seqinfo(), seqnames(), seqlevels(), seqlevels0()
seqlengths()
isCircular()
genome()
orderSeqlevels(), rankSeqlevels()
restoreSeqlevels()
sortSeqlevels(), seqlevelsInUse()
commonName(), provider(), providerVersion()
releaseDate(), bsgenomeName()
There is also a new online book, OSTA (Orchestrating Spatial Transcriptomics Analysis with Bioconductor), that collects practical, code-first material for spatial workflows. If you are building or teaching a spatial analysis pipeline, the examples and datasets there are a useful starting point.