Multiplexed spatial mapping of chromatin features, transcriptome, and proteins in tissues

Update on 10/01/2025

Uploaded the processed h5ad and seurat data for the paper into figshare(DOI: 10.6084/m9.figshare.30258697.v1).

Introduction

This repository aims to share the raw data processing and visualization code used in the "Multiplexed spatial mapping of chromatin features, transcriptome, and proteins in tissues" paper. It is now published on Nature Methods Multiplexed spatial mapping of chromatin features, transcriptome and proteins in tissues

Data analysis

1. Preprocessing the sequencing data

Next Generation Sequencing (NGS) was performed using the Illumina NovaSeq sequencer (paired-end 150 bp mode).

In the Data_preprocessing folder, directories beginning with Snakemake_* contain the code for preprocessing different modalities. The preprocessing pipeline utilizes raw FASTQ data as input, where Read 1 comprises genomic sequences, and Read 2 contains the spatial barcodes, specifically Barcode A and Barcode B.

The preprocessing pipeline we developed using the Snakemake workflow management system is in the Data_preprocessing folder. After putting the input files in the correct directory, to run the pipeline, use the command:

sbatch Snakemake.sh

Brief descriptions of the preprocessing pipeline in Snakefile:

(1) Directory and File Setup

• Automates the creation of directories for storing raw and processed data per sample.
• List samples dynamically based on the provided raw data directory.

(2) Reads Filtering

• filter_primer: Utilize bbduk to filter sequences with primers from the reads.
• filter_L1 & filter_L2: Further filtering steps target specific linker sequences.

(3) Barcode Processing

• bc_process: Extract and reformat the data. (BC_process.py)
• R1_rename: Rename and reorganize the processed reads for consistency and further processing.

(4) Barcode Tagging and Matching

• taggd: Correct barcodes and prepare the data for demultiplexing.
• add_BC: Integrate barcode information into sequencing reads based on a pre-generated list of barcodes and their matches to specific reads. (add_BC.py)

(5) Adapter Trimming

• adapter_trim: Trims sequencing adapters from the reads using Trim Galore, preparing them for alignment.

(6) Sequence Alignment

• bwa_align: Map reads to a reference genome with BWA, followed by sorting and indexing the alignments using samtools.

(7) Fragment File Generation

• fragments: Transforms BAM files into sorted, compressed, and indexed BED files using sinto toolkit, for downstream analysis.

2. Preprocessing the microscope image

Identify pixels on tissue from microscope image using Python

See the files in Image_preprocess under Data_preprocessing folder.

3. Downstream analysis

All downstream analyses were completed with R language. The package used extensively the functions in Seurat v.4.3.0.1, ArchR v1.0.2, ClusterProfiler v4.8.3, Slingshot v2.2.1, FigR v0.1.0. To ensure reproducibility, we have provided a comprehensive tutorial demonstrating the joint analysis of multiple modalities and data visualization. You can access the tutorial here: SpatialMuxSeq vignette or in Fig1_joint_analysis.Rmd. The data needed for it can be found on figshare(DOI: 10.6084/m9.figshare.27265410).

Brief descriptions of analysis scripts:

ArchR.R: Analysis of ATAC/histone modifications data.

RNA.R: Analysis of RNA data

GO.R: GO enrichment analysis for marker genes.

Multi_omics_integration.r and Multi_omics_integration_prepare.r: Multi-omics analysis.

RCTD: Cell-type annotation and deconvolution.

FigR: Gene regulation analysis

Functions:

spatial_data_visualization.R: Visualize spatially resolved data on tissue sections.

References

Pengfei Guo, Liran Mao, Yufan Chen, Chin Nien Lee, Angelysia Cardilla, Mingyao Li, Marek Bartosovic, and Yanxiang Deng. "Multiplexed spatial mapping of chromatin features, transcriptome, and proteins in tissues."