Line of best fit and regression coefficient info in pdf output for continuous data

[Marcus Leung]

Dear all

I have the following two questions regarding the maaslin pdf outputs involving continuous data:

1) Sometimes the line of best fit does not correspond to the sign of the regression coefficient (e.g. a positively-sloped fit for a negative coefficient). What are some potential explanations for that?

2) Some of the fit lines are in different colours (green vs. red vs. black). Is there a reason for that?

Thank you very much

Regards

Marcus