Imaging multi-omics: DNA seqFISH+¶

DNA seqFISH+ with sequential immunofluorescence and RNA FISH (Takei et al. 2021, 2025) measures, in the same nucleus, where thousands of loci are and what they carry: dozens of chromatin marks and nuclear bodies per spot, transcripts per cell, the cell’s position in the tissue. In a ChromData the spots carry coordinates and per-spot signals (spot_tracks), the cells carry their type, position and expression (cells, cellm), and the transcript matrix can stay in its own AnnData, linked by cell.

Step	API
read the per-FOV spot tables with the locus map and the cell clustering	`uchrom.io.read_seqfish_multiomics(csvs, locus_annotation=, cell_clustering=, out=)`
a whole replicate, with the RNA matrix as a linked AnnData	`uchrom.io.load_takei2025_cerebellum()`, `load_seqfish_multiomics_linked`
the Takei 2021 tables as FOF-CT (DNA, RNA spots, cells)	`ChromData.from_fofct(core, cell_table=, rna_table=)` (chromatin tracing)
cell positions and outlines in the tissue	`cd.cell_positions()`, `cd.set_cell_shapes()`
the full Takei 2025 cerebellum without downloading it	`uchrom.datasets.atlas("takei2025_cerebellum")`

Tutorials¶

Importing seqFISH+ multi-omics data (Takei 2025 cerebellum)

Then¶

The spots are coordinates with signals, so the common analyses apply as for chromatin tracing. Two are specific to the extra layers:

Signals along the genome: peaks of a chromatin mark in the spot signals, signals summed over intervals, annotation and sequence content — per-locus features (its tutorial calls H3K4me3 peaks on chromosome 19 of the cerebellum data).
Which cells are alike: cell embeddings from the chromatin marks (per-cell means or the mark–mark correlations within a cell) or from RNA — cells & embeddings.

The web browser shows the cells in their tissue, each cell’s traces in 3-D coloured by any mark, and the marks along the genome.

API: uchrom.io.