https://github.com/helenalc/type-state

https://github.com/helenalc/type-state

Last synced: about 1 year ago
JSON representation

• Host: GitHub
• URL: https://github.com/helenalc/type-state
• Owner: HelenaLC
• Created: 2023-04-14T08:47:23.000Z (about 3 years ago)
• Default Branch: main
• Last Pushed: 2024-11-20T10:19:30.000Z (over 1 year ago)
• Last Synced: 2025-02-17T01:29:44.683Z (over 1 year ago)
• Language: R
• Size: 298 MB
• Stars: 1
• Watchers: 2
• Forks: 0
• Open Issues: 0
• Metadata Files:
  • Readme: README.md
  • License: LICENSE

Awesome Lists containing this project

README

          
### setup

- workflow was implemented and last executed successfully with


  **R v4.4.1 with Bioc 3.20, and Python v3.11.3 with Snakemake v7.26.0**

- R version and library have to be specified in the `config.yaml` file  

  (e.g., `R: "R_LIBS_USER=/path/to/library /path/to/R/executable"`)

- `.Rprofile` is used for handling and printing command line arguments

- `logs/` capture `.Rout` files from `R CMD BATCH` executions

- `data/` contains any synthetic and real data

- intermediate results are generated in `outs/` 

- visualizations are generated in `plts/`

### workflow

- `` denotes a wildcard, namely: `t`ype, `s`tate, `b`atch,  

  `sim`ulation, `sco`re, `sel`ection, `sta`tistic,   

  `das` = differential state analysis method

- `00-get_sim/dat.R`

  - **out:** for simulations, `data/sim/00-raw/t,s,b.rds`,


    for real data, `data/dat/00-raw/.rds` (`` = dataset identifier)

  - synthetic data generation (`splatter::splatPopSimulate()`)

  - hereafter, `t,s,b` = ``


- `01-pro_sim/dat.R`

  - **in:** `data/sim|dat/00-raw/.rds`


    **out:** `data/sim|dat/01-fil/.rds`

  - minimal filtering keeping genes with count > 1  

    in ≥ 10 cells, and cells with ≥ 10 detected genes

  - log-library size normalization (`scater::logNormCounts()`)

  - highly variable gene (HVG) selection (`scran::modelGeneVar()`)

  - principal component analysis (PCA) using HVGs (`scater::runPCA()`)

- `02-sco.R`

  - **in:** `data/sim|dat/01-fil/.rds`


    **out:** `outs/sim|dat/sco-,.rds`

  - source method from one of `02-sco-.R`

  - compute gene-level metrics to quantify type-/state-specificity 

- `03-sel.R`

  - **in:** `outs/sim|dat/sco-,.rds`


    **out:** `outs/sim|dat/sel-,.rds`

  - source method from one of `03-sel-.R`

  - select genes for reprocessing

- `04-rep.R`

  - **in:** `outs/sim|dat/sco-,.rds`


    **out:** `data/sim|dat/02-rep/,.rds`

  - data reprocessing (PCA, clustering, reduction)

  

- `05-sta.R`

  - **in:** `data/sim|dat/02-rep/,.rds`


    **out:** `outs/sim|dat/sta-,,.rds`

  - source method from on of `05-sta-.R`

  - compute evaluation statistics

- `06-das.R`

  - **in:** `data/sim|dat/02-rep/,.rds`


    **out:** `outs/sim|dat/das-,,.rds`

  - source method from one of `06-das-.R`

  - perform differential state analysis (DSA)

- `07-eva.R`

  - standalone script applied to experimental data only

  - collects results across all feature selection strategies,


    selects [10, 20, ..., 90\%] for top-rank features, and recomputes


    evaluation statistics for accordingly reprocessed data (PCA, clustering)

- `08-plt_-.R`

  - **in:** `outs/sim/.rds`


    **out:** `plt/sim/-.pdf`

  - e.g., `08-plt_das-F1.pdf` collects all DSA results


    (`outs/sim/das-,,.rds`) and plots F1 scores

  - visualization of synthetic data analysis results

- `08-qlt_-.R`

  - **in:** `outs/dat/.rds`


    **out:** `plt/dat/-.pdf`

  - visualization of experimental data analysis results

  

- `09-aes.R`

  - sourced to fix the order of feature scores (`SCO`),


    ground truth-based (`DES`) and other selections (`SEL`),


    and differential state analysis methods (`DAS`) across plots

- `10-session_info.R`

  - lists and may be used to install all R packages used


    (across CRAN, GitHub, and Bioconductor), and writes the


    corresponding `sessionInfo()` output to `session_info.txt`