Variation in Mutational Robustness between Different Proteins and the Predictability of Fitness Effects.

Random mutations in genes from disparate protein classes may have different distributions of fitness effects (DFEs) depending on different structural, functional, and evolutionary constraints. We measured the fitness effects of 156 single mutations in the genes encoding AraC (transcription factor), AraD (enzyme), and AraE (transporter) used for bacterial growth on l-arabinose. Despite their different molecular functions these genes all had bimodal DFEs with most mutations either being neutral or strongly deleterious, providing a general expectation for the DFE. This contrasts with the unimodal DFEs previously obtained for ribosomal protein genes where most mutations were slightly deleterious. Based on theoretical considerations, we suggest that the 33-fold higher average mutational robustness of ribosomal proteins is due to stronger selection for reduced costs of translational and transcriptional errors. Whereas the large majority of synonymous mutations were deleterious for ribosomal proteins genes, no fitness effects could be detected for the AraCDE genes. Four mutations in AraC and AraE increased fitness, suggesting that slightly advantageous mutations make up a significant fraction of the DFE, but that they often escape detection due to the limited sensitivity of commonly used fitness assays. We show that the fitness effects of amino acid substitutions can be predicted based on evolutionary conservation, but those weakly deleterious mutations are less reliably detected. This suggests that large-effect mutations and the fraction of highly deleterious mutations can be computationally predicted, but that experiments are required to characterize the DFE close to neutrality, where many mutations ultimately fixed in a population will occur.

Design and implementation of a point-of-care computerized system for drug therapy in Stockholm metropolitan health region--Bridging the gap between knowledge and practice.

INTRODUCTION: Stockholm County Council is the largest health care provider in Sweden with an annual budget of US$ 5 billion and catering the needs of a metropolitan population of 2 million people. About 10% of health care costs are used on drugs. In 1996 Stockholm County Council decided to address the main problems associated with the process and the quality of drug prescribing. METHODS: A multiyear strategy was designed, including the establishment of a strong evidence-based organisation, Drug and Therapeutics Committees and editorial resources to adapt information to the IT-media and the development of the IT-architecture. The development and implementation of computerized tools such as a physician drug order entry system including decision support, a drug information website and electronic transmission of prescriptions were started in 1996. RESULTS: The implementation was slow at the point-of-care units. It took about 6 years before the implementation process gained speed. In September 2005 almost 1000 doctors could use the decision support system for prescribing drugs and more than 70% of all prescriptions were transmitted electronically in our region. CONCLUSIONS: The work with the strategy has shown that improvements in drug use can be accomplished by providing access to simple, rapid and safe electronic tools, but the information provided has to be associated with well-recognized regional and national expert organisations.