Back in October 2014 I put in some experimental code in that I will most likely revert because I invented a more elegant solution. This may possibly break IM and simulation files since that time although possibly not because there was some serious compatability Eric added to the robot save code. Must likely you will need the latest version for IM to work and the simulations should work.
The experimental code was added to address a problem robot called Einstein by Billy. This robot sets up some epigenetic values at the start of the simulation and then deletes the gene. This means the values where not preserved when the DNA was extracted. My solution at the time was to keep a backup of all deleted genes and 50% chance make the genes splice back in when the DNA is saved.
I just invented a cleaner solution that works much like the original Einstein. A simple gene is created to restore the epigenetic values and then is instructed to delete. This does mean it will be F3 incompatible so the feature is optional. But my current solution of restoring deleted genes will probably also be incompatible anyway.
I am hoping to start on this stuff later during the year once I get some living situation problems resolved and get access to the svn again.