Figure 1: Gene expression profiling methods and the differences between them. The raw data output format is shown for each of the six methods. Reverse transcription is not required for the two methods on the left (northern blotting and microarrays), both of which are performed directly on mRNA. Reverse transcription is required for the other four methods. The initial tissue sample can be from any source tissue or organism and can be prepared in a variety of ways (bulk tissue, cell line, etc). Three of the methods (RNA-Seq, ESTs and SAGE) utilise absolute tag counts and are therefore referred to as “digital” methods, while northern blotting, microarrays and RT-qPCR are referred to as “analogue” because they measure relative abundance levels (although they can be made absolute through use of an internal standard). The table below the figure summarises individual features of relevant techniques, provides a qualitative comparison of common expression profiling methods and compares the likelihood of errors entering the results. Detailed analysis of the individual methods are beyond the scope of this paper and we refer readers to original publications and recent reviews of these. See e.g. [138-139] for Northern blotting, [140-141] for microarrays, [142-143] for RNA-SEQ, [144-148] for PCR based methods, [149-150] for EST profiling and [151-152] for SAGE profiling methods.