• What would be the best way to display our data? Which graphics make sense?
• We want to test 1000 micro-RNAs. Which ones are associated with cancer?
• We have run high-density SNP scans for 4 studies and they have all been imputed using HapMap to a density of around 2.5 million markers. Can you run a GWAS for a 5th study (it is family-based), impute missing genotypes and then do a meta-analysis for us? We will need forest plots and QQ-plots and estimates of genomic inflation.
• The data always arrives in a format that is hard for us to use. Can you write a script for data-reformatting that we can use in future?
• If we find something interesting, we may want to show these results to regulators. Can you produce the necessary documentation and code?
• Can you write a draft manuscript for this work? We want it to be formatted for X journal.

We can do all this and more.

Examples

• We have funds to sequence 50 subjects in our study. Which ones should we do?
• We are interested in developing a companion diagnostic. Can you derive and validate a predictive model based on these candidate biomarkers? What would be the sensitivity, specificity, positive and negative predictive values? Can we have a ROC curve?
• We see a real opportunity for personalised medicine and we would like you to test this panel of biomarkers for association to drug-response. Can you produce a predictive model that would be a simple decision tree? We need one that could be easily implemented in the clinic.
• We want to investigate methylation before and after treatment with our drug. Can you tell us which genes show differential methylation?
• How many subjects do I need in order to have 80% power to detect an effect size of one standard deviation?
• We want to run a microarray experiment. Can you tell us which genes show elevated expression in diseased subjects?
• We have these XYZ measures of early clinical response to our drug. How can we combine them to predict response after 6 months? We will need to validate what we find.