Sovereign 8s.is mirror of standardgalactic/NicheTrans
  • Jupyter Notebook 98.8%
  • Python 1.2%
Find a file
Repository files (latest commit first)
Filename Latest commit message Latest commit date
2026-04-11 16:00:12 +08:00
1_Data_preparation Add files via upload 2026-04-11 15:54:48 +08:00
args Add files via upload 2026-04-11 15:54:48 +08:00
datasets Add files via upload 2026-04-11 15:54:48 +08:00
model Add files via upload 2026-04-11 15:54:48 +08:00
utils Add files via upload 2026-04-11 15:54:48 +08:00
codes_figures.txt Add files via upload 2026-04-11 15:54:48 +08:00
overall.png Add files via upload 2026-04-11 15:54:48 +08:00
README.md Add files via upload 2026-04-11 15:54:48 +08:00
requirements.txt Add files via upload 2026-04-11 15:54:48 +08:00
Tutorial_3.1__Train_NicheTrans_on_SMA_data.ipynb Add files via upload 2026-04-11 15:54:48 +08:00
Tutorial_3.2__Visualize_results.ipynb Add files via upload 2026-04-11 15:54:48 +08:00
Tutorial_3.3__Attribution_analysis.ipynb Add files via upload 2026-04-11 15:54:48 +08:00
Tutorial_4.1__Train_NicheTrans_STAR_on_SMA_data.ipynb Add files via upload 2026-04-11 15:54:48 +08:00
Tutorial_4.2__Visualize_results.ipynb Add files via upload 2026-04-11 15:54:48 +08:00
Tutorial_5.1__Train_NicheTrans_on_STARmap_PLUS_data.ipynb Add files via upload 2026-04-11 15:54:48 +08:00
Tutorial_5.2__Visualize_results_on_AD_sample.ipynb Add files via upload 2026-04-11 15:54:48 +08:00
Tutorial_5.3__Visualize_results_on_WT_sample.ipynb Add files via upload 2026-04-11 15:54:48 +08:00
Tutorial_6.1__Train_NicheTrans_STAR__on_STARmap_PLUS_data.ipynb Add files via upload 2026-04-11 15:54:48 +08:00
Tutorial_6.2__Visualize_results_on_AD_sample.ipynb Add files via upload 2026-04-11 15:54:48 +08:00
Tutorial_6.3__Visualize_results_on_WT_sample.ipynb Add files via upload 2026-04-11 15:54:48 +08:00
Tutorial_6.4__STARmap_PLUS_cell_type_attribution.ipynb Add files via upload 2026-04-11 15:54:48 +08:00
Tutorial_6.5__STARmap_PLUS_input_gene_attribution.ipynb Add files via upload 2026-04-11 15:54:48 +08:00
Tutorial_6.6__STARmap_PLUS_input_gene_attribution_niche.ipynb Add files via upload 2026-04-11 15:54:48 +08:00
Tutorial_7.1__Train_NicheTrans_on_10x_Xenium_data.ipynb Add files via upload 2026-04-11 15:54:48 +08:00
Tutorial_7.2__Visualize_results.ipynb Add files via upload 2026-04-11 15:54:48 +08:00
Tutorial_8.1__Train_NicheTrans_on_human_lymph_node_data.ipynb Add files via upload 2026-04-11 15:54:48 +08:00
Tutorial_8.2__Visualize_results.ipynb Add files via upload 2026-04-11 15:54:48 +08:00
Tutorial_9.1__Train_&_test_NicheTrans_on_MISAR-seq(atac2rna).ipynb Add files via upload 2026-04-11 15:54:48 +08:00
Tutorial_9.2__Train_&_test_NicheTrans_on_MISAR-seq(rna2atac).ipynb Add files via upload 2026-04-11 15:54:48 +08:00

NicheTrans: Spatial-aware Cross-omics Translation

This is the official Pytorch implementation of "NicheTrans: Spatial-aware Cross-omics Translation".

Overview

framework

Spatial omics technologies have revolutionized our studies on tissue architecture and cellular interactions. While spatial multi-omics offers unprecedented insights into complex biological systems, its widespread adoption is hindered by technical challenges, specialized requirements, and limited accessibility. To address these limitations, we present NicheTrans, the first spatially-aware cross-omics translation method and a flexible Transformer-based multi-modal framework. Unlike existing single-cell translation methods, NicheTrans uniquely incorporates both cellular microenvironment information and multi-modal data. We validated the advantage of NicheTrans across diverse biological cases. Through NicheTrans, we uncovered spatial multi-omics domains that were not detectable through single-omics analysis alone. Model interpretation revealed key molecular relationships, including gene programs associated with dopamine metabolism and amyloid β-associated cell states. Additionally, using translated protein markers as spatial landmarks, we quantified the spatial organization of key glial cell subtypes in the Alzheimer's Disease brain. NicheTrans represents a powerful tool for generating comprehensive spatial multi-omics insights from more accessible single-omics measurements, making multi-omics analysis more feasible for the broader research community.

Preparation

Installation

pip install -r requirements.txt
(Here, we list some packages crucial for model reproduction.)

Prepare Datasets

(1) Spatial Multimodal Analysis (SMA) dataset of Parkinson's Disease from 'Spatial multimodal analysis of transcriptomes and metabolomes in tissues'.

(2) STARmap PLUS dataset of Alizimizer's Disease from 'Integrative in situ mapping of single-cell transcriptional states and tissue histopathology in a mouse model of Alzheimer’s disease'.

(3) Human breast cancer dataset from 'High resolution mapping of the tumor microenvironment using integrated single-cell, spatial and in situ analysis'.

(4) Human lymph node dataset from 'Deciphering spatial domains from spatial multi-omics with SpatialGlue'.

(5) MISAR-seq dataset from 'Simultaneous profiling of spatial gene expression and chromatin accessibility during mouse brain development'.

The detailed processing pipelines for the SMA and STARmap PLUS datasets are located in the '1_Data_preparation' folder. Here, we released the processed spatial multi-omics data at 'GoogleDrive'. Please feel free to reach out to us if the link has expired.

Model training & testing

For each spatial multi-omics dataset, we have prepared the Jupyter files for model training and testing separately. Before model training & testing, please change the direction of the datasets in the 'args' folder.

The detailed correspondence between subfigures in the manuscript and Jupyter files is shown in:

codes_figures.txt 

Taking the SMA dataset as an example, the training process of NicheTrans is shown in:

Tutorial 3.1 Train NicheTrans on SMA data.ipynb

We provide the jupyter notebook for quantitative evaluation using pcc, spcc, and qualitative visualization of the translated results.

Tutorial 3.2 Visualize results.ipynb

Model attribution analysis

Apart from the spatial cross-omics translation, we also provided guidelines for attribution analysis in 'Tutorial 3.3', 'Tutorial 6.5', and 'Tutorial 6.6'.

'Tutorial 3.3' and 'Tutorial 6.5' are for niche-level attribution analysis across genes.

'Tutorial 6.6' can reveal intercellular cross-omics interaction in the spatial context.

Contact

If you have any questions, please don't hesitate to contact us. E-mail: zkwang00@gmail.com; zhiyuan@fudan.edu.cn.