✨ CCI-seq ✨

This repository provides the official implementation of the data analysis pipeline described in our study：

🎯 Resolving cell-cell interaction networks and their molecular logic in complex tissues

Cells in complex organisms function through extensive interactions, yet mapping these interaction networks at scale remains challenging. Here, we present CCI-seq, a high-throughput method to unbiasedly capture cell-cell interactions by combining cell clump combinatorial indexing with single-cell sequencing.

Datasets

Raw sequence reads and count matrices generated in this study are available at GSA (Genome Sequence Archive) with accession number PRJCA034558. All processed data supporting the key findings of this study are available at Zenodo repository (https://zenodo.org/records/19354406) or from the corresponding author upon reasonable request. The publicly available mouse small intestine Visium HD dataset can be accessed from the 10x Genomics website (https://www.10xgenomics.com/datasets/visium-hd-cytassist-gene-expression-libraries-of-mouse-intestine), and the spatial transcriptomics of mouse kidney are available at (Slide-seq-V2, GSE190094) and (seqFISH, https://datadryad.org/stash/dataset/doi:10.5061/dryad.bnzs7h4hj)

Requirements

• [OS] Linux (official)
• [Software]
  • seqkit: v2.5.1
  • cutadapt: v4.4
  • umi_tools: v1.1.4
  • cellranger: v7.1.0
  • Python: v3.8.5, numpy==1.23.1, pandas==1.5.1, seaborn==0.13.2, scanpy==1.9.1, scipy==1.10.1, matplotlib==3.7.5, scalex==1.0.2, sklearn==1.2.1, liana==1.5.0
  • R: v4.3.3, clusterProfiler==4.10.1

Analysis pipeline

Following sequencing, two 10x standard FASTQ libraries are generated: the mRNA library and the combinatorial index library.

For the mRNA library, we utilize the cellranger count pipeline to perform standardized preprocessing, yielding a cell-by-gene expression matrix.

cellranger count \
--id=${id} \
--sample=${sample} \
--fastqs=${fastqs} \
--transcriptome=${transcriptome} \
--maxjobs=1

For the combinatorial index library, we have established a standardized analytical pipeline from scratch, encompassing data preprocessing, quality control (QC), and cell-cell interaction analysis.

• Step 1: data preprocessing

This step involves deduplication, adapter trimming, denoising, and format standardization. It converts the input FASTQ files into a processed.txt file, the content of which is shown below:

The three columns from left to right represent barcode, UMI, and combinatorial index, respectively.

• Step 2: QC

This step aligns the mRNA library with the combinatorial index library using cell barcodes.

‼️ Note: For the 10x Genomics 3’ V3.1 Kit employed in our study, a 2-bp discrepancy exists between the corresponding barcodes of the two libraries (refer to https://zenodo.org/records/19354406 resource.zip for the mapping relationship).

Then, we perform rigorous QC and filtering on the index abundance, the rank 1 to rank 2 index ratio, and cell clump sizes—to define a unique CellID (the most abundant index) for each individual cell.

• Step 3: cell-cell interaction analysis

We use CellIDs to infer whether cells are from the same clump and introduce two metrics: Frequency and Clump to evaluate the interaction strength between cell-types. Detailed definitions of these metrics are provided in the Methods section in our paper.

We have provided analysis examples for multiple samples in the accompanying notebooks (https://github.com/zhangqf-lab/CCI-seq/tree/main/notebook) for your reference.

📍 Cite us

If you find this study useful for your research, we kindly ask that you cite our paper:

@article{XXX,
    author  = {Tang L, Tian K, Fu X, Xu Y, Wu J, Zhang J, Wang X, Ye C, Wu Q, Wu W, Feng C and Zhang QC},
    title   = {Resolving cell-cell interaction networks and their molecular logic in complex tissues},
    journal = {Nature Methods},
    year    = {2026},
    doi     = {XXX},
    url     = {XXX},
}

☎️ Contact us

For questions about the paper or code, please contact:

Kang Tian - tiankang@mail.tsinghua.edu.cn

(back to top)