It is important to indicate what statistical analysis has been performed; not just the name of the software and options selected but the method and model applied. In the case of many genes being examined simultaneously or many phenotypes a multiple comparison correction should be used to control the type I error rate or a rationale for not applying a correction must be provided. The type of correction applied should be clearly stated. It should also be clear whether the p-values reported are raw or after correction. Corrected p-values are often appropriate but raw p-values should be available in the supporting materials so that others may perform their own corrections. In large scale data exploration studies (e.g. genome wide expression studies) a clear and complete description of the replication structure must be provided.