Analysis of polysome profile experiments on B-cells.

1. Biological experiment outline:

TREATMENT

• Triplicates (2x15cm dishes = 30 milion cells per replicate), splited 5 days before 1:2
• Starved for 24 hours at 0.5 mM glucose, then 0.5 hours at 20 mM for HIGH glucose while 0.5 mM for LOW glucose

FRACTION selection for sequencing

Polysome profiling 13 fractions divided in:

• Monosomes: 5-6
• Light polysomes: 7-9
• Heavy polysomes: 10-13
• Total RNA for each replicate

2. Sequencing Analysis (Shell commands and scripts)

Raw_reads	Monosomes	Light Poly	Heavy Poly
Low_glucose rep1	10814918	9745454	12051978
Low_glucose rep2	10578664	13997704	9749807
Low_glucose rep3	13326344	11453671	7871513
High_glucose rep1	12416163	10723806	8476376
High_glucose rep2	10948574	11572776	8985906
High_glucose rep3	10855263	8762845	13196915

run_dirty_STAR.sh

• Mapping: STAR 2.5.4b; parameters: --outFilterMultimapNmax 20 --outSAMprimaryFlag AllBestScore --outSAMtype BAM SortedByCoordinate --quantMode TranscriptomeSAM GeneCounts --outStd Log --seedSearchStartLmaxOverLread 0.5 --winAnchorMultimapNmax 36 --outFilterScoreMinOverLread 0.5 --outFilterMatchNminOverLread 0.5

Unique mapped	Monosomes	Light Poly	Heavy Poly	Total RNA
Low_glucose rep1	61.24%	75.83%	74.58%	79.24%
Low_glucose rep2	56.84%	76.06%	74.69%	78.81%
Low_glucose rep3	50.12%	68.52%	73.96%	77.56%
High_glucose rep1	60.37%	73.56%	73.78%	78.71%
High_glucose rep2	58.57%	69.51%	80.24%	79.24%
High_glucose rep3	64.97%	68.10%	77.64%	78.40%

• TPM quantification: RSEM (used) and STRINGTIE (just tried)

• Filtering of the files for protein coding genes

3. Statistical analysis

diffAnal_GOenrich_Clustering_Bcells.R

R code to performe all the downstream analysis starting from the read counts and/or the TPM tables after the RNA-seq analysis.

The script is performing the following families of analysis.

1. Differential analysis using Limma.
2. Quality control using R PCA packages (and for visualisation).
3. Enrichment analysis using clusterProfiler. 4 Cluster analysis using various R clustering methods (hierarchical, k-means, mclust).

Gene ontology Analysis

...

4. RNA features analysis (Python script)

The RNA features extraction software can be found at the following GitHub repository: RNA_Features_Extraction