Example of Snakemake Workflow Implementation for Transcriptomics Tasks

Github Pipeline:

Build Status

Implemented Tasks:

1. Performed FastQC

2. Performed MultiQC

3. Trimmed Barcodes

4. Performed FastQC on Trimmed Sequences

5. Performed MultiQC on Trimmed Sequences

The biggest indicator of successful removal of adapters is the Adapter Content plot:

Before Trimming	After Trimming

6. Used STAR Aligner

7. Indexed BAM with Samtools Index, and FeatureCounts

Here we specify strandness:
0: Not stranded
1: Stranded
2: Reversely stranded

For Collibri, use 1 as it is stranded.
For KAPA, use 2 as it is reversely stranded.

8. Used DESeq2 to Perform Differential Expression (DE) Analysis Comparing UHRR vs HBR

As a result, obtained DE genes with p-adjusted values:

For Collibri:

And KAPA:

Also, volcano plots:

For Collibri:

And KAPA:

It has a Snakemake file, which will create the plot.

9. Performed PCA Using DE Genes

Plot for Collibri:

And KAPA:

It is clear that HBR and UHRR samples are separated into two distinct clusters.

More details and figures in separate deseq2.md.

10. Performed GSEA Using fgsea

We should use shrunken fold changes.

Performing on Reactome pathways did not give promising results, so performed on KEGG pathways.

Collibri showed slightly better results (with smaller p-adjusted values) than KAPA but found similar pathways. Some pathways were related to cancer:

• KEGG_BASAL_CELL_CARCINOMA
• KEGG_COLORECTAL_CANCER
• KEGG_ENDOMETRIAL_CANCER
• KEGG_WNT_SIGNALING_PATHWAY
• KEGG_PATHWAYS_IN_CANCER
• KEGG_REGULATION_OF_ACTIN_CYTOSKELETON

For Collibri:

For KAPA:

DAG of Snakemake Workflow: