Principal Investigator David Gifford
We are developing methods to take advantage of correlations within and between high-throughtput sequencing experiments. This work formalizes the notion that the Poisson distribution, commonly used in sequencing data analysis, does not fit real world sequencing data well. Instead of suggesting that people use some type of more complicated distribution, we developed a method which can preprocess and re-weight data so that existing Poisson based pipelines work correctly.