scMMT 1.0.1

A package for cell annotation, protein prediction, and low dimensional embedding representation

scMMT

scMMT (single-cell multi-modal data and multi-task learning tool) is a powerful deep learning computational tool designed for the analysis of CITE-seq and scRNA-seq data. It offers various functionalities such as cell annotation, protein expression prediction, and low-dimensional embedding. With scMMT, researchers can efficiently explore and interpret complex single-cell datasets, enabling deeper insights into cellular heterogeneity and intercellular interactions.

Create environment

conda create -n scMMT python=3.10
conda activate scMMT

Installation

pip install scMMT

Alternatively, you can also install the package directly from GitHub via

pip install git+https://github.com/SongqiZhou/scMMT.git

Example Demo:

Guided Tutorial

Data

(1) Seurat 4 human peripheral blood mononuclear cells (GEO: GSE164378).

(2) H1N1 influenza PBMC dataset (https://doi.org/10.35092/yhjc.c.4753772).

(3) COVID dataset(https://www.ebi.ac.uk/arrayexpress/experiments/E-MTAB-10026/).

The University of Pennsylvania has put these data sets together for the convenience of downloading. Download Here. The reference github link is: https://github.com/jlakkis/sciPENN_codes

Software Requirements

• Python >= 3.10
• torch >= 2.0.0
• scanpy >= 1.9.3
• scikit-learn >= 1.2.2
• scikit-learn-intelex >= 2023.1.1
• pandas >= 2.0.1
• numpy >= 1.24.3
• scipy >= 1.10.1
• tqdm >= 4.65.0
• anndata >= 0.9.1