Search

The cost of drug development is increasing, and investment returns are decreasing. The number of drugs approved by FDA is in decline in terms of the number of new molecular entities (NMEs). Amongst the reasons noted for this are the adverse side effects and reduced efficiency of many of the potential compounds. This is a problem both for the pharmaceutical industry and for those suffering from diseases for which there are no or few available treatments. Advances in computational chemistry, computer science, structural biology and molecular biology have all contributed to improved drug design strategies and reduced the time taken for drug discovery. By interfacing cheminformatics and bioinformatics with systems biology we can create a powerful tool for understanding the mechanisms of patho-physiological systems and identifying lead molecules for various diseases. This integration of drug design approaches can also play a crucial role in the prediction and rationalization of drug effects and side effects, improving safety and efficacy and leading to better approval rates. Addressing the lack of knowledge on the fundamental aspects of the various computational tools for drug discovery, this book is a compilation of recent bioinformatics and cheminformatics approaches, and their integration with systems biology. Written primarily for researchers and academics in chem- and bioinformatics, it may also be a useful resource for advanced-level students.

On-chip systems (also known as System-on-chip or soc) are more and more complex. soc design heavily relies on reuse of building blocks, called ips (Intellectual Property). These ips are built by different designers working with different tools. So, there is an urgent demand for interoperability of ips, that is, ensuring format compatibility and unique interpretation of the descriptions. ip-xact is a de facto standard defined in the context of electronic system design to provide portable representations of (electronic) components and ips. It succeeds in syntactic compatibility but neglects the behavioral aspects. uml is a classical modeling language for software engineering. It provides several model elements to cover all aspects of a design (structural and behavioral). We advocate a conjoint use of uml and ip-xact to achieve the required interoperability. More specifically, we reuse the uml Profile for marte to extend uml elements with specific features for embedded and real-time systems. marte Generic Resource Modeling (grm) package is extended to add ip-xact structural features. marte Time Model extends the untimed uml with an abstract concept of time, adequate to model at the Electronic System Level. The first contribution of this thesis is the definition of an ip-xact domain model. This domain model is used to build a uml Profile for ip-xact that reuses, as much as possible, marte stereotypes and defines new ones only when required. A model transformation has been implemented in atl to use uml graphical editors as front-ends for the specification of ips and to generate ipxact code. The second contribution addresses the modeling of the ip time properties and constraints. uml behavioral diagrams are enriched with logical clocks and clock constraints using the marte Clock Constraint Specification Language (ccsl). The ccsl specification can serve as a \golden model" for the expected time behavior and the verification of candidate implementations at different abstraction levels (rtl or tlm). Time properties are verified through the use of a dedicated library of observers.