1 of 10

Workshop 2

Data analysis workflows

​

​

Martina Summer-Kutmon, PhD

9 September 2021

​

2 of 10

Agenda

  • Explanation setup / tools

​

  • Transcriptomics data (RNAseq / microarray)
    • Workflow 1 - Pathway enrichment
    • Workflow 2a - PPI network

​

  • NanoString data / proteomics data
    • Workflow 2b- PPI network
    • Workflow 3 - Network diffusion analysis

​

​

3 of 10

Setup/Tools

4 of 10

Setup / Tools

  • Implemented in R
  • Currently workflows available as RMarkdown files

​

  • Explanations, code and questions for interpretation

​

  • Run code blocks iteratively

​

  • Discuss results

  • Feedback workflow / visualizations / requests?

​

​

5 of 10

Transcriptomics data analysis workflows

    • Workflow 1 - Pathway enrichment
    • Workflow 2a - PPI network

6 of 10

Workflow 1 - Pathway enrichment

  • Input: differential gene expression (log2FC, pvalue for all genes!)

​

  • Example dataset: classical and basal pancreatic cancer subtype compared to healthy

​

  • Steps
  • DEG visualization
  • Pathway enrichment
  • Pathway visualization
  • Pathway crosstalk
  • Drug-target extension

​

​

Aim

Find interesting pathways altered in the dataset, study the molecular mechanisms known and investigate pathway crosstalk.

7 of 10

Workflow 2a - PPI network

  • Input: differential gene expression (log2FC, pvalue)

​

  • Example dataset: classical and basal pancreatic cancer subtype compared to healthy

​

  • Steps
  • PPI network of differentially expressed genes
  • Extend with pathway information

​

​

Aim

Many DEGs not present in pathways -> study their functional relationships in PPI networks. Extend with pathway information to see what is already covered.

Identify knowledge gaps.

8 of 10

NanoString data / proteomics data analysis workflows

    • Workflow 2b- PPI network
    • Workflow 3 - Network diffusion analysis

​

9 of 10

Workflow 2b - PPI network

  • Input: differential expression (log2FC, pvalue)

​

  • Example dataset: NanoString Folfirinox treatment

​

  • Steps
  • PPI network of differentially expressed genes
  • Extend with pathway information
  • Visualize data

​

​

Aim

Study expression changes taking functional relations into account.

Study pathway associations.

10 of 10

Workflow 3 - Network diffusion analysis

​

  • Input: differential gene expression (log2FC, pvalue)

​

  • Example dataset: NanoString Folfirinox treatment

​

  • Steps
  • Pathway-Gene association network
  • Identify starter nodes (DEGs)
  • Run heat diffusion algorithm
  • Identify highest ranked pathways

​

​

Aim

Find interesting pathways in the dataset (direct involvement of the immune genes measured and more distant)