Biomedical Genomics Core

Designing your experiment to optimally answer the biological question of interest requires careful planning. Through consultation with the Core staff, we will guide you through this process, help you choose the right platform, and give advice on how to prepare your samples. Experimental design depends on the study goal, but these general guidelines may help with this process:

• Simplicity. Keep the design of your experiment as simple as possible. If this is your first attempt at a microarray analysis experiment do not be tempted to include multiple conditions or time points. Rather, if it is possible, just compare two conditions (e.g. treated vs. untreated or KO vs. WT) or focus on the single time point that would likely answer the question you are most interested in. The more complex the design of your experiment, the greater is the potential for being overwhelmed with the output produced.
• Replicates. The number of biological replicates (do not consider a technical replicate the same as a true biological replicate) used in any given study is directly proportional to the confidence with which one can call any given gene on the array as being differentially expressed. Where possible avoid pooling samples, as although this will reduce variation, there still needs to be enough replicate assays in order to perform statistics. We require an absolute minimum of four biological replicates.
• Confounding Factors. Array analysis is very sensitive to the introduction of confounding factors in your experimental design. For example, hybridization date is an important factor in microarray analysis so where possible we aim to process all samples together. Similarly when performing your experiment aim to randomize confounding factors that cannot be avoided, such as RNA extraction date or sample harvest date. Avoid preparing samples from different groups at different times and avoid the introduction of gender bias by comparing a test group of males with a control group of females.