Table of contents

• Clinical data
  • Sample ID
  • Patient characteristics
  • MSI and TMB
  • Tumor stage
  • Survival data
  • Immune cell estimation from blood
  • Tumor-Infiltrating Lymphocytes (TIL) estimation from images
  • Coverage
  • Immune cell estimation from ESTIMATE algorithm
• STAD clinical data
• UCEC clinical data
• Neoantigens from pVACseq
• Mucin Gene Expression

Clinical data

Sample ID

• Tumor_Sample_Barcode, case_submitter_id, sample_submitter_id: These are unique codes assigned to each sample from the TCGA (The Cancer Genome Atlas) dataset. For more information look at GDC
• dataset: Specifies the type of TCGA dataset. Since COAD (colon adenocarcinoma) and READ (rectal adenocarcinoma) datasets are combined, this column tracks which dataset each sample belongs to.
• sample_type: Indicates the type of sample. TCGA includes different sample types such as normal solid tissue and primary tumor. Since we focus only on primary tumors, this column is mainly for verification.

Patient characteristics

• age: The age of the patient at the time of diagnosis, represented as years to birth.
• gender: The gender of the patient.

MSI and TMB

• MSI_TMB: A classification combining Microsatellite Instability (MSI) status and Tumor Mutational Burden (TMB). This classification includes:

  • "MSI-H" (Microsatellite Instability-High)
  • "MSS/TMB-H" (Microsatellite Stable with High Tumor Mutational Burden)
  • "MSS/TMB-L" (Microsatellite Stable with Low Tumor Mutational Burden)

  These are the three main groups studied in this research. For further details, refer to my draft (email me if you dont have access at xkdn@zhaw.ch).

• MSI: Provides specific information on the microsatellite instability status.

• TMB: Represents the tumor mutational burden, a measure of the number of mutations per megabase in the tumor genome.

Tumor stage

• pathologic_stage: Overall classification of the cancer stage based on tumor size, lymph node involvement, and metastasis.
• pathology_T_stage: Specifies the size and extent of the primary tumor.
• pathology_N_stage: Indicates the involvement of nearby lymph nodes.
• pathology_M_stage: Describes whether the cancer has metastasized to distant parts of the body.

Survival data

• residual_tumor: The amount of remaining tumor after surgery, categorized as R0 (no residual tumor), R1 (microscopic residual), or R2 (macroscopic residual).
• overall_survival: The time (in days) from diagnosis until death or last follow-up.
• alive (OS): Binary indicator of patient survival status (1 = alive, 0 = deceased).

Immune cell estimation from blood

These columns represent different types of immune cells measured in the tumor microenvironment:

• lymphocytes: A type of white blood cell crucial for the immune response.
• neutrophils: White blood cells involved in inflammation and response to infection.
• eosinophils: A type of white blood cell associated with allergic reactions and certain infections.
• mast_cells: Immune cells that play a role in allergic responses and inflammation.
• dendritic_cells: Antigen-presenting cells that help trigger immune responses.
• macrophages: Large immune cells that engulf and digest cellular debris and pathogens.

Tumor-Infiltrating Lymphocytes (TIL) estimation from images

• number_of_TIL_patches: The count of TIL patches identified in tumor images.
• number_of_clusters: The number of immune cell clusters detected in the tumor.
• til_percentage: The estimated percentage of tumor-infiltrating lymphocytes based on image analysis.
• leukocyte_fraction: The fraction of leukocytes (white blood cells) present in the tumor microenvironment.
• immune_subtype: Classification of tumors based on immune cell composition and activity.
• global_pattern: A descriptor of immune infiltration patterns in tumor tissues.

Data is taken from this study:

Saltz, J., Gupta, R., Hou, L., Kurc, T., Singh, P., Nguyen, V., Samaras, D., Shroyer, K.R., Zhao, T., Batiste, R., et al.: Spatial organization and molecular correlation of tumor-infiltrating lymphocytes using deep learning on pathology images. Cell reports 23(1), 181–193 (2018)

Coverage

• mean_coverage: The average depth of sequencing coverage across the genome.
• proportion_coverage_30x: The proportion of the genome covered at a minimum of 30x sequencing depth.

Immune cell estimation from ESTIMATE algorithm

The ESTIMATE algorithm uses gene expression data to infer different tumor properties:

• stromal: Measures the presence of stromal cells in the tumor microenvironment.
• immune: Estimates the level of immune cell infiltration in the tumor.
• estimate: Overall score combining stromal and immune components.
• purity: An estimation of tumor purity, i.e., the proportion of cancer cells in the sample.

ESTIMATE algorithm paper:

Yoshihara, K., et al.: Inferring tumour purity and stromal and immune cell admix-ture from expression data. Nat Commun 4, 2612 (2013) https://doi.org/10.1038/ncomms3612

STAD clinical data

stad_clinical.csv file contains clinical annotations for samples from the TCGA-STAD (Stomach Adenocarcinoma) dataset.

The dataset includes information such as sample and case identifiers, sample type, age, gender, tumor stage, MSI status, tumor mutational burden (TMB), overall survival, and vital status. It follows the same column structure as the COAD and READ clinical datasets (see the clinical data section for reference).

If certain columns are missing, it is because the information was not available in the original STAD clinical data source.

UCEC clinical data

ucec_clinical.csv file contains clinical annotations for samples from the TCGA-UCEC (Uterine Corpus Endometrial Carcinoma) dataset. It follows the same column structure as the COAD and READ clinical datasets (see the clinical data section for reference).

If certain columns are missing, it is because the information was not available in the original UCEC clinical data source.

Neoantigens from pVACseq

This dataset was generated using pVACseq, a tool that predicts neoantigens based on sequencing data. Neoantigens are mutations in tumor cells that can be recognized by the immune system and are crucial in immunotherapy response prediction.

• Tumor_Sample_Barcode: Unique identifier for each sample in the TCGA dataset.
• variant_type: Specifies the type of variant found in the neoantigen analysis. Include missense mutations, frameshift (FS) mutations and inframe deletion (inframe_del). An inframe deletion is a type of mutation where one or more codons (sets of three nucleotides) are deleted from the DNA sequence without disrupting the overall reading frame of the gene. .
• neoantigens: The number of neoantigens detected for each variant type in the sample.

Mucin Gene Expression

We have included a new dataset, muc_tpm.csv, which contains TPM-normalized gene expression values for mucin-related genes. This allows comparison of mucin expression levels across samples.