Skip to content

Sorennorge/MacAulayLab-RNAseq3-Zucker

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

9 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

MacAulayLab - Obese Zucker rats with testosterone treatment

The work and scripts are done by the MacAulay Lab.
All programs used are free and open-source. In the interest of open science and reproducibility, all data and source code used in our research is provided here.
Feel free to copy and use code, but please cite:
(coming soon)
Remember rewrite file_names and folder_names suitable for your pipeline.

The RNAseq and Analysis follows these steps:

Raw data analysis - Library Build, Mapping and Quantification

The analysis uses RNA STAR for mapping and RSEM for TPM quantification.

RNA-STAR and RSEM Library build and indexing

0.1.1 - RNA_STAR_Indexing.sh
0.2.1 - RSEM_Indexing.sh

RNA-STAR Mapping and RSEM quantification

0.1.2 -RNA_STAR_Analysis.sh
0.2.2 - RSEM_Analysis.sh

Zucker rats (Obese vs lean)

Create count tables and reduce for RNA star

1.1.1 - Raw data GeneCounts.py

Create count tables and reduce for RSEM (TPM)

1.2.1 - Raw data RSEM.py

Differential expression analysis with DEseq2

2.1.1 - DE Analysis Zucker Obese.R

Volcano plot

2.2.1 - DE Analysis Volcano.R

Heatmap plot

2.3.1 - DE Analysis Heatmap.R

Piechart of differentially expressed genes in percentage

2.4.1 - Pie chart - Differentially expressed genes.py

Update lookup DE tables

3.1.1 - Update lookup DE tables.py

Go term enrichement analysis - protein coding genes

4.1.1 - Enrichment - Convert raw Panther DB data.py
4.1.2 - Enrichment - Plot.py

Transport analysis

5.1.1 - Enrichment - Transport analysis - Convert raw data.py
5.1.2 - Enrichment - Transport analysis - subcellular location.py
5.1.3 - Enrichment - Transport analysis - Generate count tables.py

Testosterone treated Zucker rats (Obese TT vs lean (VEH))

Create count tables and reduce for RNA star

8.1.1 - Raw data - GeneCounts.py

Create count tables and reduce for RSEM (TPM)

8.2.1 - Raw data - RSEM.py

Differential expression analysis with DEseq2

9.1.1 - DE Analysis - Zucker Obese TT.R

Volcano plot

9.2.1 - DE Analysis - Volcano - Obese TT.R

Heatmap plot

9.3.1 - DE Analysis - Heatmap - Obese TT.R

Piechart of differentially expressed genes in percentage

9.4.1 - Pie chart - Differentially expressed genes.py

Update lookup DE tables

10.1.1 - Update lookup DE tables - Obese TT.py

Go term enrichement analysis - protein coding genes

11.1.1 - Enrichment - Convert raw Panther DB data.py
11.1.2 - Enrichment - Plot.py

Transport analysis

12.1.1 - Enrichment - Transport analysis - Convert raw data.py
12.1.2 - Enrichment - Transport analysis - subcellular location.py
12.1.3 - Enrichment - Transport analysis - Generate count tables.py

Calculate adjusted p-value for transport

13.1.1 - Add adjusted pvalue to transporter tables.py

Create supplementary tables

14.1.1 - Create supplementary tables

About

No description, website, or topics provided.

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published