From the Principal Component Analysis, it is clear liver cancer samples can be distinguished from breast cancer samples as well as from non-tumorous liver samples based on transcriptomics profiles of samples. Next, we can find specific genes that are representative of the found differences. Since we have 2 factors at play (Tissue type and tumorous vs. non-tumorous), we can perform factor regression analysis to find which genes can be associated with differences between cancer-normal and liver-breast.

Link for Input data:

Quantile Normalized data for 20530 genes: https://raw.githubusercontent.com/pine-bio-support/Final_Liver_cancer_project/main/LIHC-BRCA_QN5_data_Pipeline_LIHC_BRCA_20531_genes_FPKM_values_data_Threshold_5_normalized.txt

Refer the steps to run the factor analysis pipeline on the T-Bioinfo Server as mentioned in the coursework: https://learn.omicslogic.com/Learn/project-03-tcga-liver-cancer-precision-oncology/lesson/02-pan-cancer-analysis

The results obtained by running the pipeline includes p values to differentiate conditions based on chi square. 

For any questions or access to the T-Bioinfo server, you can reach out to us on support@pine.bio