I have two treatment groups with 4 biological replicates each. I measured 100 lipid species in each of them and want to visualize the differences using a volcano plot.

Which of these two ways is the correct to process my data:

• Calculate for each lipid species the averages, fold changes, adjusted p-values between the two treatment groups and then at the end log2 transform the fold changes and -log10 transform the p-values for plotting
• Log2 transform all the measured lipid concentrations first, calculate the averages, fold changes, adjusted p-values between the two treatment groups. Plot the fold changes without further transformation and -log10 transform the p-values for plotting