rodogi / rna-seq

Pipeline for RNA-seq data

IDEA

Unified pipeline for RNA-seq data procesing:

Warning

All scripts should be executed from RNA-seq (main directory).

• e.g. $ Rscript ./R/ReadData.R

Pre-conditions

• The raw data files are already downloaded from GDC.
• The raw data is organized accorgding to the Data Structure section.
• The raw data has been cuantified using htseq for hg38.pXX Human genome.
• R required packages according to respective section.

Post-conditions

• Obtain the normalized expression matrix.
• The data is stored in a Summarized Experiment object.

Data Structure

The data follows this directory structure

The data follows this directory structure
|--data
|     | - raw
|     |     |- metadata.txt
|     |     |-> manifest
|     |     |    |- experimentalcondition1.manifest
|     |     |    |- experimentalcondition1.manifest
|     |     |    ...
|     |     |    |- experimentalconditionN.manifest
|     |     |-> experimentalcondition1
|     |     |    |- ID_bla-bla.htseq.counts
|     |     |-> experimentalcondition2
|     |     |    |- ID_bla-bla.htseq.counts
|     |     |...
|     |     |-> experimentalconditionN
|     |          |- ID_bla-bla.htseq.counts
|     | - summarized_experiment                  
|           |- Out.RData
|--R
   |- ReadData.R

• The md5.txt contains the md5 data for each file.
• The metadata.txt file contains all metadata for each sample.
• The Manifest folder contains the respective sample metadata in a tab separeted format.
• The Experimental-Condition-X folder contains the respective htseq.counts data files. For example:

  Ensembl_ID.Version \t Raw_Counts
  ENSG000000003.13      25

• The Summarized_Experiment folder contains the RData object.
• The R directory contains the different R files.

R

requirements.R

• Check if the required R and Bioconductor packages are installed
• Install the R and Bioconductor missing packages

ReadData.R

1. Read Experimental Data
2. Read Biomart Data
3. Merge count and annotations

Conventions

• Variable and function names: camel case.
• Column names: lower case and underscore notation.
• Experimental condition names: lower case without underscore.
• Folder names: lower case with underscore.
• 80 characters per line.
• Function definition: above.
• Source code file name: lower case without underscore.