I am going to assess relationship between respiratory disease and environmental factors using OLS regression. The result of the OLS analysis showed a strong adjusted R squired value (0.74) with statistically significant coefficients and no redundancy among variables. The residuals are not spatially auto correlated. However the residuals are not normally distributed. Both Jarque-Bera and Koenker BP statistics are statistically significant. I tried several samples with different variables the result was same. Can you please suggest me a solution?