MLOmics: Cancer Multi-Omics Database for Machine Learning

Abstract

Framing the investigation of diverse cancers as a machine learning problem has recently shown significant potential in multi-omics analysis and cancer research. Empowering these successful machine learning models are the high-quality training datasets with sufficient data volume and adequate preprocessing. However, while there exist several public data portals including The Cancer Genome Atlas (TCGA) multi-omics initiative or open-bases such as the LinkedOmics, these databases are not off-the-shelf for existing machine learning models. We propose MLOmics, an open cancer multi-omics database aiming at serving better the development and evaluation of bioinformatics and machine learning models. MLOmics contains 8,314 patient samples covering all 32 cancer types with four omics types, stratified features, and extensive baselines. Complementary support for downstream analysis and bio-knowledge linking are also included to support interdisciplinary analysis.

Data Records

The MLOmic main datasets are now accessible under the Creative Commons 4.0 Attribution (CC-BY-4.0), which supports its use for educational and research purposes. The main datasets presented include all cancer multi-omics datasets corresponding to the various tasks.

Main Datasets

The Main Datasets repository is stored as csv files and is organized into three layers. The first layer contains three files corresponding to three different tasks: Classification_datasets, Clustering_datasets, and Imputation_datasets. The second layer includes files for specific tasks, such as GS-BRCA, ACC, and Imp-BRCA. The third layer contains three files corresponding to different feature scales, i.e., Original, Aligned, and Top. The omics data from different omics sources are stored in the following files: Cancer_miRNA_Feature.csv, Cancer_mRNA_Feature.csv, Cancer_CNV_Feature.csv, and Cancer_Methy_Feature.csv. Here, Cancer represents the cancer type, and Feature indicates the feature scale type.
The ground truth labels are provided in the file Cancer_label_num.csv, where Cancer represents the cancer type.
The patient survival records are stored in the file Cancer_survival.csv.

Code Availability

All code are stored on the cloud in user-friendly formats and are accessible via our GitHub repository (https://github.com/chenzRG/Cancer-Multi-Omics-Benchmark). We provide comprehensive guidelines for utilization. All files are ready for direct loading and analysis using standard Python data packages like Numpy or Pandas. We hope it can lower the barriers to entry for machine learning researchers interested in developing methods for cancer multi-omics data analysis, thereby encouraging rapid progress in the field.