作者: Junfeng Chen , Magnus Rattray
关键词: Position bias 、 Sequence-tagged site 、 Computational biology 、 Massively parallel signature sequencing 、 DNA microarray 、 Bias (Epidemiology) 、 Positive correlation 、 Biology 、 Genetics 、 Gene expression profiling
摘要: Massively Parallel Signature Sequencing (MPSS) technology was recently developed as a high-throughput for measuring the concentration of mRNA transcripts in sample. It has previously been observed that position signature tag transcript (distance from 3' end) can affect measurement, but this effect not studied detail. We quantify tag-position bias Classic and MPSS using published data Arabidopsis, rice human. investigate relationship between measured nonlinear regression methods. The is shown to be broadly consistent across different sets. find there exist significant biases both data. For data, genes with middle-range have highest abundance on average while high-range, far end, show decrease. high-range tend flatter abundance. Thus, our results confirm method fixes substantial problem method. positive correlation low-range genes. Compared effects length number exons, seems more Arabadopsis. reflected count proportion unexpressed identified. Tag-position should taken into consideration when technology,