I have mutation data of several genes and want to determine the association among their mutation status (variable has only two values 0 for not mutated and 1 for mutated for all the genes. I wish to present all the associations in a single figure/ table so I wish to avoid using Fisher's exact test. Any suggestions are welcomed.
If I simply make a correlation matrix for it, would it be acceptable statistics?
Kendall rank correlation is considered better than Spearman's correlation. Pearson's is not appropriate for this case (it is for quantitative variables).
Mohit Arora Dr. Milanesi is absolutely right. Follow the discussion on the link (Is it meaningful to calculate Pearson or Spearman correlation between two Boolean vectors?)
As far as your data will have a lot of 0s, I´d try using 0 and 1000 as NO/YES values to emphasize mutations. May be it´s a silly trick but may be it helps with decimal issues.
- Badges
- Science topic
- Similar topics
- Engineering
- Petroleum Engineering
- Oil