Amplification-based expression data show bias towards long gene misregulation in synaptic disorders. Homo sapiens

Several recent studies have suggested that genes that are over 100 kb in length are particularly likely to be misregulated in neurological diseases associated with synaptic dysfunction, such as autism, Fragile X syndrome, and Rett syndrome. These length-dependent transcriptional changes seem to be modest, but, given the low sensitivity of high-throughput transcriptome profiling technology, the statistical significance of these results needs to be reevaluated. Here we show that transcriptional changes reflected in microarray and RNA-Sequencing benchmark datasets from the SEQC Consortium show a bias toward genes of greater length, even in the comparison of technical replicates. We hypothesized that PCR amplification, which is used in both microarray and RNA-Seq technologies, could be introducing this bias. We found that, when the fold-change values are small, PCR amplification in microarray and RNA-Seq technologies does produce a bias toward longer genes; we found no similar bias with nCounter technology, which is not based on PCR amplification. We provide an approach to more rigorously assess length-dependent changes that begins with comparing randomized control samples to estimate baseline gene length dependency and evaluate the statistical significance of gene length regulation. Overall design: We purchased Universal Human Reference RNA from Agilent Technologies, Inc., and Human Brain Reference RNA from Life Technologies, Inc. For the nCounter experiments, we used the same RNA sample types as SEQC. We assessed RNA purity and integrity with Bioanalyzer (Agilent Technologies, Inc.) prior to use in the nCounter assays. Sample preparation and analysis were done using an nCounter Prep Station 5s and an nCounter Digital Analyzer 5s.