Investigation of factors affecting RNA-seq gene expression calls

Abstract

RNA-seq enables quantification of the human transcriptome. Estimation of gene expression is a fundamental issue in the analysis of RNA-seq data. However, there is an inherent ambiguity in distinguishing between genes with very low expression and experimental or transcriptional noise. We conducted an exploratory investigation of some factors that may affect gene expression calls. We observed that the distribution of reads that map to exonic, intronic, and intergenic regions are distinct. These distributions may provide useful insights into the behavior of gene expression noise. Moreover, we observed that these distributions are qualitatively similar between two sequence mapping algorithms. Finally, we examined the relationship between gene length and gene expression calls, and observed that they are correlated. This preliminary investigation is important for RNA-seq gene expression analysis because it may lead to more effective algorithms for distinguishing between true gene expression and experimental or transcriptional noise.

Effect of low-expression gene filtering on detection of differentially expressed genes in RNA-seq data

“…Intergenic mapping distributions may be used to quantify RNA-seq experimental noise [7]. We quantify RNA-seq “noise” expression by aligning reads to intergenic regions, which we define as regions complementary to genic regions (including two flanking sequences of lkb at both ends) annotated by Aceview.…”

“…In addition, we investigate other methods for determining filtering thresholds, including percentiles of intergenic distribution [7] and LODR (limit of detection ratio) introduced in erecdashboard [8]. LODR is derived from the analysis of external spike-in RNA control ratio mixtures, and is defined as the minimum count above which a gene with an absolute log fold-change signal has a 100% chance of obtaining a statistically significant adjusted p-value [8].…”

“…Other methods for determining filtering thresholds include estimation based on the distribution of intergenic reads per kilobase per million mapped reads (RPKM) values [7] and estimation based on External RNA Controls Consortium (ERCC) spike-in controls [8]. However, there is no systematic comparison of different filtering methods, and no investigation of how different RNA-seq pipelines affect these filtering methods.…”

Direct nanopore sequencing of mRNA reveals landscape of transcript isoforms in apicomplexan parasites

“…This is vital because the transcripts encode altered or truncated proteins which may exhibit deleterious activity [84]. Some studies postulate that the predicted alternative splice events are the result of either experimental or transcriptional noise [85], or that a substantial portion of such transcripts are contaminating pre-mRNA molecules, and so do not represent true alternative splicing [86,87]. Nevertheless, many RNA-seq-based analyses operate on the assumption that PTC transcripts are biologically significant or relevant [88].…”

Covering all your bases: incorporating intron signal from RNA-seq data

“…However, it has been shown that intron reads account for a significant proportion of sequencing reads (10-12) yet it is not common practice to quantify these reads. Perhaps this is due to suggestions that intron reads represent experimental and transcriptional noise (13), or are unusable in exon and gene quantification (14).…”