Joint radar-communications (JRC) is considered to be a vital technology in deploying the next generation systems, since its useful in decongestion of the radio frequency (RF) spectrum and utilising the same hardware resources for dual functions. Using JRC systems for dual function generates interference between both the operations which needs to be addressed in future standardization. Furthermore, JRC systems can be advanced by deploying hybrid beamforming which implements fewer number of RF chains than the number of transmit antennas. This paper designs a robust hybrid beamformer for minimizing the interference of a JRC transmitter via RF chain selection resulting into mutual information maximization. We consider a weighted mutual information for the dual function JRC system and implement a common analog beamformer for both the operations. The mutual information maximization problem is formulated which is non-convex and difficult to solve. The problem is simplified to convex form and solved using Dinkelbach approximation abased fractional programming. The performance of the optimal RF selection based proposed approach is evaluated, compared with baselines and its effectiveness is inferred via numerical results.