2008-2009 Session
October 8th 2008, 2pm to 5pm at MANDEC (Manchester Dental Education Centre), Higher Cambridge Street (tea will be served about mid-afternoon) (building 41, entrance on corner facing building 35) Joint meeting with Manchester University's Biostats Group Theme: Analysis challenges arising from post-genomic high-dimensional datasets XIAYI KE (Centre for Integrated Genomic Medical Research, School of Medicine, University of Manchester)Genome-wide association studies of rheumatoid arthritis Genome wide association studies have been very successful in identification of disease susceptibility loci for various complex human diseases in recent 2-3 years. The Wellcome Trust Case Control Consortium (WTCCC) is the best example of this development, where seven common complex diseases, including rheumatoid arthritis (RA), have been investigated and a number of disease susceptibility loci have been uncovered. Before genome-wide association was applied to RA, there were only two genetic loci known to be associated with RA. Now more than ten novel loci have been identified. Despite these successes, all these genetic factors identified so far can only explain about 10% of the disease susceptibility variance, leaving wide open where is the rest of genetic contribution and how to find each element. In this talk, RA will also be used as an example to illustrate the classical relationships between sample size, power, effect size, and type I errors. Challenges in localization of causal variants and detection of gene-gene interactions will also be discussed. DAVID HOYLE (Health Methodology Research Group, School of Community Based Medicine, University of Manchester)Properties of sample covariance matrices from modern high-dimensional genomic data sets Modern biotechnology is providing a wealth of new data sets that are high-dimensional but comparatively low sample size. This presents challenges for traditional methods of analysis. Understanding the properties of sample covariance matrices under such conditions has become a recent research focus. In this talk I will outline new findings and my own research in this area. Neil Lawrence (School of Computer Science, University of Manchester)
Meeting contact: Stephan Rudolfer 10th December, 5pm, Room G.107, Alan Turing Building, University of Manchester. Complex intervention trials should be able to answer both pragmatic and explanatory questions in order to test the theories motivating the intervention and help understand the underlying nature of the clinical problem being tested. Key to this is the estimation of direct effects of treatment and indirect effects acting through intermediate variables, such as mediators. Using psychological treatment trials as an example of complex interventions, this talk explains statistical methods which evaluate both direct and indirect effects in the presence of hidden confounding between mediator and outcome. We introduce principal stratification and structural mean models, and discuss approaches for attaining identifiability of key parameters of the basic causal model. Assuming that there is no direct effect of treatment leads to the use of instrumental variable methods, using randomisation and its interactions with baseline covariates as instruments. The new methodology is illustrated with motivating examples of randomised trials from the mental health literature. Meeting contact: Stephan Rudolfer 15th January, 2009 All Day on Statistics in the City Meeting contact: Margaret Martin February 11th, 5pm, Room G.107, Alan Turing Building, University of Manchester. Adaptive designs gained considerable popularity in recent years due to their increased use in clinical research. They provide flexibility and efficiency in managing studies without compromising the scientific methods used. For example, safety and efficacy data can be reviewed during scheduled interim analyses and the results can be taken into account to modify the trial methodology. Typical examples of such modifications are early stopping of a trial, dropping an arm or modifying the sample size. In this talk the methodology of the adaptive designs will be discussed and opportunities for using the optimal design theory to make them highly efficient will be illustrated. Possibilities for using adaptive designs in preclinical trials will also be demonstrated. Meeting contact: Stephan Rudolfer March 11th, 5pm, Room G.107, Alan Turing Building, University of Manchester. Consider the situation where someone has broken into a house. The perpetrator forced the door open and stole valuables from the premises. A neighbour saw the perpetrator trying to open the door by forcing the door handle. The police, using a cotton swab, retrieved organic material from the door handle and a DNA profile was obtained from the swab. After a tip off, the police arrested Mr X and a reference profile has been obtained from him. The profile obtained from the door handle and the suspect, Mr X, are similar. The courts in the UK expect a number that summarises the evidential value of the DNA findings. In this talk the following topics will be discussed:
Meeting contact: Stephan Rudolfer 29th April, 5pm, Room G.107, Alan Turing Building, University of Manchester. Meeting contact: John Logsdon July 8th 2009, 4.30pm, Room G.107, Alan Turing Building, University of Manchester
|