Home

Dates: 22-25.05.2023

Conceptual dependencies

Stuff that we'd need to have covered in previous courses (Rintro, Rdeseq2):

• ggplot2 & tidyverse
  • patchwork, with the conceptual idea of infix operator represented as plus sign being defined for different classes...
• Futureverse's future.apply() with plan 'multicore' (relevant seurat vignette)
• Working with S4 classes & methods (we'll use a lot of these: seurat-object)
• Theoretical knowledge?
  • Dimensional reduction (PCA, tSNE, UMAP)
  • Clustering (JackStraw, KNN, k-means)
  • Enrichment, and its variations. (Odds ratio, GSEA, Pathway Analysis.)

To Decide

• dataset - 1 or 2? 10X ideally
  • 10x PBMC ‘4k’ and ‘4k_small’ (both bundled with Seurat)
    • official site with other Ks
  • Integration (ATAC, Spatial, etc)
    • See: SingleCellMultiModal, this package has several vignettes, each dedicated to a different dataset
    • Another one: WeberDivechaLCdata
• Are we covering any external functionality? see wrappers

Outline

Day1 - Adrian

• Import data
  • Read10X
• Basic QC
  • iSEE and/ or scCustomize
  • shiny brushedplot
• Normalization
• Clustering + cluster QC

Day2 - Katarzyna

• Embedding
• Integration (batch effects correction)
  • Harmony, kBET
  • sctransform
  • maybe it's worth generating a pooled dataset like this

Day3

• Cell annotation
  • See this
• Differential Expression
  • Use some rather high threshold to speed-up https://www.singlecellcourse.org/scrna-seq-dataset-integration.html

Day4

• Pseudotime and trajectory analysis

---

Left out?

• Spatial transcriptomics
• Velocity