Compare groups

What to do: Define two patient subgroups and ask K-Pro to compare them across a specific data type. Be explicit about the outcome and the groups.

For survival comparisons, K-Pro reports Cox proportional-hazards results by default: hazard ratios with 95% confidence intervals and p-values, adjusted for multiple comparisons when there are more than two groups. It falls back to a log-rank test only if the Cox model cannot be fitted. For other comparisons, such as expression between two groups, you can name the test you want (for example a t-test or Wilcoxon test).

Example prompt:

In the TCGA-LUAD cohort, test whether there is a statistically significant difference in overall survival between patients with and without a KRAS mutation. Report the hazard ratio, 95% confidence interval, and p-value.

Expected result: A Kaplan-Meier curve for the two groups, with the Cox hazard ratio, 95% CI, and p-value.

Follow-up:

Run a multivariate Cox regression adjusting for age, tumor stage, and smoking status to check whether KRAS mutation status is an independent prognostic factor. Show the result as a forest plot.

Expected result: A multivariate forest plot showing the adjusted hazard ratio and 95% CI for each covariate.