I have a dataset containing Newt presence$-$absence variable $(0-1),$ fish $($which are the newt main predator$)$ presence$-$absence variable $(0-1),$ and a list of continuous variables such as many compounds and elements $($compounds known to lead to eutrophication, some ions, some heavy$-$metals$),$ and some categorical variables such as water clarity, conductivity, alkalinity.

The study aims to understand which of these variables is considered a pollutant, if the pollutant affect the newt's presence in these ponds and if the fish presence impacts the newt's presence. I found correlation between a lot of these variables. so did a PCA to decrease multicollinearity. And with a glm I found that the next variables are linked to pollution in ponds that affect negatively the newt's presence negatively : alkalinity, water clarity class, NPOC, As$,$ Cu$,$ Cd$,$ Be$,$ BrO$3.$

The final question is : $"$Is a negative effect of fish on newt occurrence stronger in polluted ponds than in less polluted ponds?" i used a Mixed model $($GLMER$).$ do you think it's the best model for this questiin? could you give me the exact code? with this model i found out that it has no singificance$.$ however I have a problem: i want to do a scatterplot where on the X axis i put waterclarity gradient, and on the plot we see the points corresponding to the Fish presence and others corresponding to the newt's presence $($using $2$ different colors to differentiate the $2$ variables$).$ on this scatter plot i would like to draw a line of best fit for each of the $2$ variables, the fish presence and the newt presence. could you help me with the code to do this scatterplot. i also question if i should trasform my presence$-$absence data $($newt and fish$),$ or standardize it to do this scatterplot?