Unlike today's reactive approaches, information flow based approaches can provide positive assurances about overall system integrity, and hence can defend against sophisticated malware. However, there hasn't been much success in applying information flow based techniques to desktop systems running modern COTS operating systems. This is, in part, due to the fact that a strict application of information flow policy can break existing applications and OS services. Another important factor is the difficulty of policy development, which requires us to specify integrity labels for hundreds of thousands of objects on the system. This paper develops a new approach for proactive integrity protection that overcomes these challenges by decoupling integrity labels from access policies. We then develop an analysis that can largely automate the generation of integrity labels and policies that preserve the usability of applications in most cases. Evaluation of our prototype implementation on a Linux desktop distribution shows that it does not break or inconvenience the use of most applications, while stopping a variety of sophisticated malware attacks.