RNAseq_analysis
- Purpose:
WARNING: betta version. You can explore the code, but it is a very early version.
Useful pipeline parts and functions for RNAseq analysis.
RNAseq tecnhologies:
| Type | Description | Expected Features (#Raw reads,corr,#variables in final test) |
|---|---|---|
| Conventional RNAseq | mRNA sequencing. Classic | 13-30 million reads. 80-95% Pearson's correlantion by pair of samples after normalization. Low-expressed tags ~ 3-8k genes |
| Single-Cell RNAseq | mRNA seq from a single cell | 2-8 million reads. 90% Pearson's correlation by pairs. ~ #13k low-expressed tags. |
| small RNAseq | Tipically microRNAs sequencing | 4-15 million reads, depending on soMir analysis. Unknown but many zeros as Single-cell. |
ToDo List
Quality Control:
- Plot : Cross-species contamination plot by using the Fast_screen output. Babraham Institute http://www.bioinformatics.babraham.ac.uk/projects/fastq_screen/
- Plot : Improved FastQC reports somehow. Babraham Institute http://www.bioinformatics.babraham.ac.uk/projects/fastqc/
- Plot : Snapshot of alignment given a Exon ID or Gene ID (EnsEMBL, UCSC, genenames, etc).
Data exploration:
- R function : Principal Component Analysis.
Differential gene expression:
- R function : Batch effect removal from gene expression matrix by Generalized Linear Models.
- R function : limma-voom implementation.
- R function : Gene Set test statistics.
- Plot : Gene Set test barcode plot improvements: worm and Enrichment Score.
Gene Set Testing:
- Perl script : Mapping Human -> Mouse | Mouse -> Human gene names based on MGI.
- R function : Pre-ranked file (.rnk) generator.
Overall dependencies
Fastq_screen(perl script) - Babraham InstituteRSeQC(python scripts) - [Broad Institute] (http://rseqc.sourceforge.net/)edgeR(R/Bioconductor)limma(R/Bioconductor)ggplot2(R)GSEABase(R/Bioconductor)lattice(R)
References
Wang L, Wang S, Li W. RSeQC: quality control of RNA-seq experiments Bioinformatics (2012) 28 (16): 2184-2185. doi: 10.1093/bioinformatics/bts356