SINGLE-CELL-RECITER

Single-cell topics modeling using deep learning and multi-omics enrichment analysis.

Abstract

In this project, we will apply the amortized Latent Dirichlet Allocation (LDA) model to scRNA. Initially developed in the natural language processing field, LDA is a method of modeling topics using a topic matrix. We can apply the method to single-cell biology by treating each cell as a document and each gene expression count as a word. After training the model, we will plot the topics over a UMAP of the reference set and inspect them for characteristic gene sets. Multiomics enrichment analysis will be performed further using ranked gene/word lists of each cell cluster/document.

Table of Contents

• SINGLE-CELL-RECITER
  • Single-cell topics modeling using deep learning and multi-omics enrichment analysis.
  • Abstract
  • Table of Contents
  • Background
  • Workflow
  • Data
  • Usage
    • 👩‍💻 Installation
      • Git
    • Requirements
    • Create and activate 🐍 conda environment
    • Steps to run
      • Step 1
      • Step 2
      • Step 3
      • Step 4
      • Step 5
  • !Results
  • Team Members
  • Mentor

Background

Natural language processing introduced Latent Dirichlet Allocation (LDA) as a topic modeling method. We can apply the method to single-cell biology by treating each cell as a document and each gene expression as a word. We can apply the method to single-cell biology by treating each cell as a document and each gene expression as a word.

Workflow

• Pancheva A, Wheadon H, Rogers S, Otto TD. Using topic modeling to detect cellular crosstalk in scRNA-seq. PLoS computational biology. 2022 Apr 8;18(4):e1009975.

Data

SRA Run Selector

Run	BioProject	BioSample	LibraryLayout	Organism	Sample Name	source_name	tissue
SRR11832836	PRJNA603103	SAMN13919403	PAIRED	Homo sapiens	GSM4284223	Skin	cSCC
SRR11832837	PRJNA603103	SAMN13919401	PAIRED	Homo sapiens	GSM4284224	Skin	Normal Skin
SRR11832838	PRJNA603103	SAMN13919399	PAIRED	Homo sapiens	GSM4284225	Skin	cSCC
SRR11832839	PRJNA603103	SAMN13919396	PAIRED	Homo sapiens	GSM4284226	Skin	Normal Skin
SRR11832840	PRJNA603103	SAMN13919395	PAIRED	Homo sapiens	GSM4284227	Skin	cSCC
SRR11832841	PRJNA603103	SAMN13919394	PAIRED	Homo sapiens	GSM4284228	Skin	Normal Skin
SRR11832842	PRJNA603103	SAMN13919393	PAIRED	Homo sapiens	GSM4284229	Skin	cSCC
SRR11832843	PRJNA603103	SAMN13919392	PAIRED	Homo sapiens	GSM4284230	Skin	Normal Skin
SRR11832844	PRJNA603103	SAMN13919391	PAIRED	Homo sapiens	GSM4284231	Skin	cSCC
SRR11832845	PRJNA603103	SAMN13919390	PAIRED	Homo sapiens	GSM4284232	Skin	cSCC
SRR11832846	PRJNA603103	SAMN13919389	PAIRED	Homo sapiens	GSM4284233	Skin	Normal Skin
SRR11832847	PRJNA603103	SAMN13919388	PAIRED	Homo sapiens	GSM4284234	Skin	cSCC
SRR11832848	PRJNA603103	SAMN13919387	PAIRED	Homo sapiens	GSM4284235	Skin	Normal Skin
SRR11832850	PRJNA603103	SAMN13919385	PAIRED	Homo sapiens	GSM4284237	Skin	Normal Skin
SRR11832851	PRJNA603103	SAMN13919384	PAIRED	Homo sapiens	GSM4284238	Skin	cSCC
SRR11832852	PRJNA603103	SAMN13919383	PAIRED	Homo sapiens	GSM4284239	Skin	Normal Skin
SRR11832853	PRJNA603103	SAMN13919382	PAIRED	Homo sapiens	GSM4284240	Skin	cSCC
SRR11832854	PRJNA603103	SAMN13919381	PAIRED	Homo sapiens	GSM4284241	Skin	Normal Skin
SRR11832855	PRJNA603103	SAMN13919380	PAIRED	Homo sapiens	GSM4284242	Skin	cSCC
SRR11832856	PRJNA603103	SAMN13919379	PAIRED	Homo sapiens	GSM4284243	Skin	Normal Skin
SRR11832857	PRJNA603103	SAMN13919378	PAIRED	Homo sapiens	GSM4284244	Skin	cSCC
SRR11832858	PRJNA603103	SAMN13919377	PAIRED	Homo sapiens	GSM4284245	Skin	Normal Skin
SRR11832859	PRJNA603103	SAMN13919434	PAIRED	Homo sapiens	GSM4284246	Skin	cSCC
SRR11832860	PRJNA603103	SAMN13919433	PAIRED	Homo sapiens	GSM4284247	Skin	Normal Skin
SRR11832849	PRJNA603103	SAMN13919386	PAIRED	Homo sapiens	GSM4284236	Skin	cSCC

Usage

👩‍💻 Installation

Git

To fetch source code, change in to directory of your choice and run:

git clone -b main \
    git@github.com:u-brite/Single-cell-reciter.git

Requirements

OS:

Tested on:

Description:	🎩 Red Hat Enterprise Linux Server release 7.9 (Maipo)
Release:	7.9
Codename:	Maipo

Tools:

• Anaconda3
  • Tested with version: conda 4.11.0
• Cellranger
  • cellranger-5.0.1
• bamtofastq v1.3.5

Create and activate 🐍 conda environment

❗ Optional: Depends on project. ❗

Change in to root directory and run the commands below:

# create conda environment. Needed only the first time.
conda env create -f Hackin_Omics.yml

# if you need to update existing environment
conda env update --file Hackin_Omics.yml

# activate conda environment
conda activate Hackin_Omics

Steps to run

Step 1

Use the public scRNA data or your own dataset. Various public datasets are available from a variety of sources, including NCBI GEO.

Step 2

Using the Cell Ranger is a set of analysis pipelines to align reads, and generate feature-barcode matrices.

Example:

#!/bin/bash
#
#SBATCH --job-name=CellRanger_P1
#SBATCH --output=CellRanger_P1
#SBATCH --ntasks=1
#SBATCH --cpus-per-task=20
#SBATCH --partition=largemem
#SBATCH --time=12:00:00
#SBATCH --mem-per-cpu=10000

module load CellRanger/5.0.1
cellranger -V

time cellranger count --id=P1_1 \
--fastqs=P1_Fastaqs \
--sample=SRR11832837,SRR11832839 \
--transcriptome=refdata-gex-GRCh38-2020-A

Once all the Cell ranger analysis is done move all *.h5 file to one folder as per the need.

Note: Assuming HCP is being used for the analysis.

Step 3

As described in the scRNA_Topic_Modeling.ipynb load the *.h5 file and perform the topic modeling analysis.

Output from this step includes -

Results/
|-- Rank_by_topic.csv
|-- SCC.h5ad

Step 4

Using the each topic in the matrix from the previous step perfom the multiomics gene ranked encriment analyis as discribed in the NOTEBOOK Topic_Ranked_Enrichment_Analysis.ipynb.

Step 5

(Future work) Using the topic in the matrix from the step 3 and perform co-expression analysis. see Notebook: 5_co_expression_analysis.ipynb

Team Members

Name	Email	Role
👋 Nilesh Kumar (Ph.D. Candidate)	nileshkr@uab.edu	Team Leader
🤚 Dr. Virginie Grosboillot	virginie.grosboillot@alumni.ethz.ch	Team Member
✋ Hammad Ali Hassan	hammadali50@live.com	Team Member

Mentor

Dr. Shahid Mukhtar Associate Professor, Department of Biology Co-Director, Genetics and Genomic Sciences Undergraduate Program Faculty, Department of Surgery Scientist, Nutrition Obesity Research Center UAB | The University of Alabama at Birmingham Campbell Hall 369 | 1300 Univ. Blvd. | Birmingham, AL, 35294-1170