Flexynesis manuscript material

Publication material relevant for the manuscript describing the flexynesis software package.

Our manuscript currently available at BioRxiv.

See our github repository of Flexynesis.

Datasets used in the manuscript

Publicly available datasets

• CCLE.rds: downloaded from Zenodo.
• GDSC2.rds: downloaded from Zenodo.
• lgggbm_tcga_pub.tar.gz: downloaded from cbioportal.
• brca_metabric.tar.gz: downloaded from cbioportal.
• depmap: downloaded from depmap portal.
• nbl_target_2018_pub.tar.gz: downloaded from cbioportal.
• GDCData: TCGA cohort datasets for 33 cancer types downloaded using the TCGABiolinks package (See GitHub).
• prot-trans: protein sequence embeddings obtained from prot-trans-xl-uniref50 model on uniprot sequences.
• describeProt: protein level sequence/structure/function features from describeprot database (Download here).
• coadread_tcga_pan_can_atlas_2018.tar.gz: downloaded from cbioportal.
• brca_tcga_pan_can_atlas_2018.tar.gz: downloaded from cbioportal.
• gbm_tcga_pan_can_atlas_2018.tar.gz: downloaded from cbioportal.

PREPARED datasets used as input to Flexynesis

The datasets listed above were further processed to create train/test splits for training using Flexynesis. The prepared datasets and the output of Flexynesis model training can be downloaded from Zenodo archive: https://zenodo.org/records/16442998

The ./prepared folder contains:

• ccle_vs_gdsc: Drug response data from cell lines from CCLE and GDSC2 datasets.
• lgggbm_tcga_pub_processed: Merged cohorts of LGG + GBM samples.
• brca_metabric_processed: METABRIC dataset processed.
• single_cell_bonemarrow: CITE-Seq dataset from Seurat.
• tcga_vs_ccle: TCGA tumors and CCLE cell lines from 3 different cancer types: lung cancer, glioma, and breast cancer
• tcga_cancertype: TCGA cancer cohort for ~21 cancer types 100 samples per each cohort.
• depmap_gene_dependency: Dataset for gene-dependency prediction in cell lines. Consists of depmap gene expression + prottrans embeddings + describeprot features.
• panGI_msi: Gene expression and promoter methylation data from 7 different TCGA cohorts (gastrointestinal and gynocological cancers) with microsatellite instability (MSI) annotations: TCGA-COAD (Colon Adenocarcinoma), TCGA-ESCA (Esophageal Carcinoma), TCGA-PAAD (Pancreatic Adenocarcinoma), TCGA-READ (Rectum Adenocarcinoma), TCGA-STAD (Stomach Adenocarcinoma), TCGA-UCEC (Uterine Corpus Endometrial Carcinoma), TCGA-UCS (Uterine Carcinosarcoma).

Flexynesis output for use-cases

For the different use-cases described in the manuscript, Flexynesis output (along with the configurations used) can be downloaded from here: https://bimsbstatic.mdc-berlin.de/akalin/buyar/flexynesis_manuscript_material/manuscript_processed_data.tgz

Environment

Clone the manuscript repo:

git clone https://github.com/BIMSBbioinfo/flexynesis_manuscript.git

Install flexynesis

mamba create -n flexynesisenv python==3.11 snakemake
mamba activate flexynesisenv
pip install flexynesis

Install other packages

guix package --manifest=guix.scm --profile=./manuscript

Activate environment

source ./manuscript/etc/profile
mamba activate flexynesisenv

Figures

Assuming the prepared datasets and Flexynesis output files are downloaded from the following locations:

• Datasets: https://zenodo.org/records/16442998/files/datasets_prepared.tgz
• Flexynesis output: https://zenodo.org/records/16442998/files/manuscript_processed_data.tgz

The figures in the manuscript can be reproduced using the following instructions:

Unzip the Flexynesis datasets and output folders:

tar -xzvf manuscript_processed_data.tgz
tar -xzvf datasets_prepare.tgz

Activate guix environment:

source ./flexynesis_manuscript/manuscript/etc/profile

Change to folder with Flexynesis output data

cd manuscript_processed_data

Figure 1: Flexynesis workflow

This figure was manually made.

Figure 2: single-task figures

Rscript ../flexynesis_manuscript/src/figure2.R ../flexynesis_manuscript/src/utils.R single_multi_experiments panGI_MSI_analysis/output

Figures 3 and 4: multi-task figures

Rscript ../flexynesis_manuscript/src/figure3_and_4.R ../flexynesis_manuscript/src/utils.R single_multi_experiments

Figure 5: unsupervised clustering (TCGA cancer types)

Rscript ../flexynesis_manuscript/src/figure5.R ../flexynesis_manuscript/src/utils.R ./unsupervised_cancertype/

Figure 6: cross-modality prediction of cell line dependency probabilities

Rscript ../flexynesis_manuscript/src/figure6.R ../datasets_prepared/depmap_gene_dependency/ depmap_analysis/output/

Figure 7: demonstration of fine-tuning

Rscript ../flexynesis_manuscript/src/figure7.R ../flexynesis_manuscript/src/utils.R finetuning/

Figure 8 and Supp. Figure 9: marker analysis

Rscript ../flexynesis_manuscript/src/figure8.R ../flexynesis_manuscript/src/utils.R marker_analysis/

Figure 9: benchmark summary

Rscript ../flexynesis_manuscript/src/figure9.R benchmarks/output

Supp Figure 10: Run Times & Resources

Rscript ../flexynesis_manuscript/src/supp_figure_10.R runtimes/output

Supp Figure: Run Times & Resources

Rscript ../flexynesis_manuscript/src/figures_runtimes.R runtimes/output

Collate all figure source data files

Rscript ../flexynesis_manuscript/src/collate_figure_source_data.R ../flexynesis_manuscript/data/Figure_Source_Data/