1. Data overview
Petr V. Nazarov, LIH
2023-10-24
0. Installations
Please, install R and RStudio. For RStudio you will need admin-rights, but R you could install in your home dir.
1.1. RNA-seq Data Generation
Short dictionary
- read: short fragment detected by RNA-seq
- library: collection of all reads from the sample
- CPM: counts per million nucleotides
- TPM: transcripts per million (proportion)
- FPKM: fragments per kilobase of exon per million reads mapped
- RPKM: reads per kilobase of exon per million reads mapped. (for single-end)
10-minute simple explanation of TPM / FPKM is here.
1.3. Sequence-based QC: FastQC
FastQC – a simple but widely-used Java-based tool for quality control of the experiments at the sequence level. It provides a modular set of analyses which you can use to give a quick impression of whether your data has any problems of which you should be aware before doing any further analysis.
See Examples.
You can combine FastQC results into an interactive dashboard usin MultiQC. Short introduction in this video.
1.5. Data Repositories
- GEO - largest repository with free omics data (mainly transcriptomics).
- ArrayExpress - EU-based analogue of GEO.
- TCGA - ~ 11k cancer samples studied by various omics techniques.
- GTEx - ~ 17k normal (post mortem) samples studied by RNA-seq.
Take home messages 1
- RNA-seq can be used as row counts and normalized (TPM, FPKM). See what you need for a specific algorithm!
- For QC of your samples at sequence level – use FastQC. To combine results - MultiQC
- Expression-related data in transcriptomics are strongly right-skewed. Therefore:
for statistics use either precise distribution (negative binomial for RNA-seq) or work with log-transformed data
use log-transformed data for exploratory analysis and visualization
- Several large repositories of the data exist. Before planning your experiments – make a search for existing data
| Home Next |
By Petr Nazarov
|
