Lab instructions

Clone the course repository

Retrieve the entire repository including all datasets and notebooks.

cd ~/Desktop/course
git clone git@github.com:NBISweden/workshop_omics_integration.git .

Inside you will find the following folders:

• environments/ - all conda environments necessary for running notebooks
• session_preprocessing/ - data pre-processing, prior to course start
• session_ml/ - the machine learning sessions, days 1 and 2
• session_topology/ - the network topology analysis, day 3 morning
• session_meta/ - network meta-analysis, day 3 afternoon
• session_nmf/ - session on matrix factorization, SNF and recommender systems, day 4 morning
• session_gems/ - the genome-scale modeling session, day 4 afternoon
• session_gsa/ - gene set and reporter feature, day 5 morning
• .../ - remaining folders not necessary for running any of the contents

You will need to create specific conda environments as indicated below.

Environments See the pre-course installation, specifically 3. Create and activate the environment.
Environments - we merged the environments for many notebooks down to 4 environments.

• Data pre-processing (linux: env-preprocessing_linux.yaml, macOS: env-preprocessing.yaml)

• Supervised Integration & Feature selection& (linux: env-ml_linux.yaml

  macOS: env-ml.yaml)

• Meta analysis (linux

  macOS /session_meta/renv.lock)

• All remaining notebooks (linux env-ml_nets_linux.yaml

  MacOS env-ml_nets.yaml)

Alternatively, you can find smaller environments below.

Before the course: data pre-processing

• Data pre-processing:
  • environment macOS: /environments/env-preprocessing.yaml
  • environment linux: /environments/env-preprocessing_linux.yaml
  • notebook (html): /session_preprocessing/preprocessing.ipynb

Day 1: Supervised and Unsupervised integration

In MacOSX you need to install XQuartz if you get errors related with rgl:

install.packages('session_ml/rgl_0.100.54.tgz', repos = NULL, type="source", dependencies = TRUE)

• Supervised integration:
  • environment macOS: /environments/env-ml.yaml
  • environment linux: /environments/env-ml_linux.yaml
  • notebook (html): /session_ml/SupervisedOMICsIntegration/supervised_omics_integr_CLL/supervised_omics_integr_CLL.Rmd
• Feature selection:
  • environment macOS: /environments/env-ml.yaml
  • environment linux: /environments/env-ml_linux.yaml
  • notebook (html): session_ml/FeatureSelectionIntegrOMICs/OmicsIntegration_FeatureSelection.Rmd
• Unsupervised integration:
  • environment macOS: /environments/env-ml_day2.yaml
  • environment linux: /environments/env-ml_day2_linux.yaml
  • notebook (html): session_ml/UnsupervisedOMICsIntegration/UnsupervisedOMICsIntegration.Rmd

Day 2: Dimensionality reduction, deep learning, single-cell omics integration

After activating the environment as above, open the .Rmd below by launching Rstudio with rstudio & or the jupyter notebooks with jupyter-notebook.

• Dimensionality reduction:
  • environment macOS: /environments/env-ml_day2.yaml
  • environment linux: /environments/env-ml_day2_linux.yaml
  • notebook (html): session_ml/DimReductSingleCell/OmicsIntegration_DimensionReduction.Rmd
• UMAP for data integration:
  • Note that the data needs to be decompressed before running the notebook gunzip scRNAseq.csv.gz; gunzip scProteomics.csv.gz
  • environment macOS: /environments/env-ml_day2.yaml
  • environment linux: /environments/env-ml_day2_linux.yaml
  • notebook (html): session_ml/UMAP_DataIntegration/UMAP_DataIntegration.ipynb
• Deep learning:
  • environment macOS: /environments/env-ml_day2.yaml
  • environment linux: /environments/env-ml_day2_linux.yaml
  • notebook (html): session_ml/DeepLearningDataIntegration/DeepLearningDataIntegration.ipynb
• Single cell:
  • environment macOS: /environments/env-ml_day2.yaml
  • environment linux: /environments/env-ml_day2_linux.yaml
  • notebook (html): session_ml/SingleCell/SingleCell_OmicsIntegration.Rmd

Day 3: Network topology

• Network topology:
  • environment macOS: /environments/env-merged_nets.yaml
  • environment linux: /environments/env-merged_nets_linux.yaml
  • notebook (html): /session_topology/lab.ipynb
• Meta analyses:
  • environment macOS:/session_meta/renv.lock
  • environment linux:/session_meta/renv.lock
  • notebook (html): /session_meta/lab_meta-analayses-v2.Rmd

Day 4: GEM structure and simulation with Cobrapy

• Genome-scale metabolic modeling:
  • environment macOS: /environments/env-merged_nets.yaml
  • environment linux: /environments/env-merged_nets_linux.yaml
  • notebook (html): /session_gems/lab.ipynb

Day 5: Gene set and reporter analysis

• Gene set analysis and reporter features:
  • environment macOS: /environments/env-merged_nets.yaml
  • environment linux: /environments/env-merged_nets_linux.yaml
  • notebook: (html): /session_gsa/GEM_GSA.Rmd