p-values histogram for random effects

I have recently come across with this post: http://varianceexplained.org/statistics/interpreting-pvalue-histogram/ it was great to understand better what my results mean, but I still have a remaining question I can seam to find the answer for.
To give you some context, I work with DNA methylation data investigating what changes might occur in response to exercise. My interest is manly at the response at the individual level. To be able to investigate consistent changes in DNA methylation at the individual level, we have performed a repeated testing intervention, where we collected samples for analyses multiple timepoints during the intervention (baseline, 4-week, 8-week, 12-week) this way we can build individual slopes and observe if participants show a consistent change or not. My model for this was as it follows:
lmer(DNAmethylation ~ Timepoint + Age + (1+Timepoint|ID), data=data)

Timepoints were coded as continuous (1,2,3,4)

The fixed effect of this model gives me the information for changes at the group level that are associated with DNA methylation, while the random effect gives me information at the individual level. After obtaining my results I have created a histogram of p-values (See attached). The histogram for the fixed effect looks fine and I have an anti-conservative p-values, for those I can just keep going with my analyses and correct for multiple testing right?
But for the random effects I get the Conservative p-values which in your post you imply that something may be wrong with my test. I just don't know what could it be as I have been using this model based on other publications and advice from a bioinformatician. Could it be that p-values histograms are just not fit for random effects? What would you do in this situation?