Classifying Cancers of unknown primary

• Collect data htseq of mRNA Breast and Colon Data.
• KNN Impute zeroes
• Linearise about mean of 1
• binary classifier SVM (linear kernel, gaussian otherwise)
• Multiclass classifier for more cancer types.
• Colon vs Colorectal
• CNV and MEth
• Other data types multi kernel.

Build up a dataset of breast and prostate expression data for training Design an api to collect data and preprocess for future purposes

Start with 2 cancers and do classification of each (Breast and Prostate) with expression data

Order of number of samples GDC is Lung (4977), Breast (3682), Colorectal (2755) (of which is colon (2305) and rectum (371)), Kidney (2201). Try using Breast and colon sample size 200 of each.

[https://portal.gdc.cancer.gov/]