Severe case of overfitting in my research

I'm a MSc student in bioinformatics. What I do is I gather transcriptomic data from many cancer datasets, I conduct some analysis over each dataset sepratly, get important cells and genes as features, and use them in a machine learning model to predict a target variable.
The analysis in which I get the cells scores is pretty solid. It is based on the transcriptomic data, and it basically tells me how much is there from each cell type in each …

artificial case overfitting research