Dataset Information

Measure transcript integrity using RNA-seq data

ABSTRACT: Achieved biospecimens annotated with patient clinical characteristics are unique resources for translational research. However, RNA extracted from the achieved tissues is often degraded. RNA degradation can have a significant impact on the measure of transcript abundance that can lead to an increase rate of erroneous differentially expressed genes. Here, we are presenting the transcript integrity number (TIN) algorithm to measure the RNA degradation at transcript level. When applied to RNA-seq datasets generated from human brain Glioblastome cell line, human peripheral blood mononuclear cells, and metastatic castration resistant prostate cancer (mCRPC) clinical tissues, TIN provided a more reliable and more sensitive measure of RNA degradation than RIN, as demonstrated by much higher concordance with the RNA fragment size estimated from read pairs. More importantly, when comparing 10 mCRPC samples with lower RNA quality to another 10 samples with higher RNA quality, we demonstrated that calibrating gene quantification with TIN scores could mitigate RNA degradation effects and greatly improve gene expression analysis. The detected differentially expressed genes before TIN correction were predominantly ribosomal genes. However, when we adjusted gene quantifications with the corresponding TIN scores, we found differentially expressed genes were highly enriched in prostate cancer specific pathways. When further evaluating the performance of TIN correction using synthetic spike-in transcripts with predetermined abundance in RNA-seq data generated from Sequencing Control Consortium (SEQC), we found TIN adjustment had a better control of false positives and false negatives (sensitivity = 0.89, specificity = 0.91), as compared to gene expression analysis results without TIN correction (sensitivity =0.98, specificity = 0.50).

ORGANISM(S): Homo sapiens

PROVIDER: GSE70285 | GEO | 2016/02/19

SECONDARY ACCESSION(S): PRJNA288159

REPOSITORIES: GEO

ACCESS DATA

Similar Datasets

Project description:Achieved biospecimens annotated with patient clinical characteristics are unique resources for translational research. However, RNA extracted from the achieved tissues is often degraded. RNA degradation can have a significant impact on the measure of transcript abundance that can lead to an increase rate of erroneous differentially expressed genes. Here, we are presenting the transcript integrity number (TIN) algorithm to measure the RNA degradation at transcript level. When applied to RNA-seq datasets generated from human brain Glioblastome cell line, human peripheral blood mononuclear cells, and metastatic castration resistant prostate cancer (mCRPC) clinical tissues, TIN provided a more reliable and more sensitive measure of RNA degradation than RIN, as demonstrated by much higher concordance with the RNA fragment size estimated from read pairs. More importantly, when comparing 10 mCRPC samples with lower RNA quality to another 10 samples with higher RNA quality, we demonstrated that calibrating gene quantification with TIN scores could mitigate RNA degradation effects and greatly improve gene expression analysis. The detected differentially expressed genes before TIN correction were predominantly ribosomal genes. However, when we adjusted gene quantifications with the corresponding TIN scores, we found differentially expressed genes were highly enriched in prostate cancer specific pathways. When further evaluating the performance of TIN correction using synthetic spike-in transcripts with predetermined abundance in RNA-seq data generated from Sequencing Control Consortium (SEQC), we found TIN adjustment had a better control of false positives and false negatives (sensitivity = 0.89, specificity = 0.91), as compared to gene expression analysis results without TIN correction (sensitivity =0.98, specificity = 0.50). RNA sequencing of 20 bone-metastatic castration resistant prostate cancer (mCRPC) using Illumina HiSeq 2500. Out of 20 mCRPC samples, 10 samples have relative low RNA integrity and another 10 samples have relative higher RNA integrity as measured by Agilent RIN score.

Project description:While the importance of random sequencing errors decreases at higher DNA or RNA sequencing depths, systematic sequencing errors (SSEs) dominate at high sequencing depths and can be difficult to distinguish from biological variants. These SSEs can cause base quality scores to underestimate the probability of error at certain genomic positions, resulting in false positive variant calls, particularly in mixtures such as samples with RNA editing, tumors, circulating tumor cells, bacteria, mitochondrial heteroplasmy, or pooled DNA. Most algorithms proposed for correction of SSEs require a training data set, which is typically either from a part of the data set being M-bM-^@M-^\recalibratedM-bM-^@M-^] (Genome Analysis ToolKit, or GATK) or from a separate data set with special characteristics (SysCall). Here, we combine the advantages of these approaches by adding synthetic RNA spike-in standards to human RNA, and use GATK to recalibrate base quality scores with reads mapped to the spike-in standards. Compared to conventional GATK recalibration that uses reads mapped to the genome, spike-ins improve the accuracy of Illumina base quality scores by a mean of 5 units, and by as much as 13 units M-BM- at CpG sites. In addition, since reads mapping to the genome are not used for recalibration, our method allows run-specific recalibration even for the many species without a comprehensive and accurate SNP database. We also use GATK with the spike-in standards to demonstrate that the Illumina RNA sequencing runs overestimate quality scores for AC, CC, GC, GG, and TC dinucleotides, while SOLiD has less dinucleotide SSEs but more SSEs for certain cycles. We conclude that using these DNA and RNA spike-in standards with GATK improves base quality score recalibration. Four human RNA samples with equimolar ERCC spike-in standards were sequenced on Illumina. Two human brain/liver/muscle RNA mixtures with dynamic range of ERCC spike-in standards were sequenced on SOLiD.

Project description:Experiments conducted on this tiling array are used to (1) validate the frozen gene sets of the current genome annotation, (2) improve the predicted gene structures by empirically determining UTRs and intron-exon boundaries, identifying missing upstream, internal, and downstream exons and alternative transcripts, (3) propose gene structure models in transcribed regions containing no predicted genes and (4) delineate transcriptionally active regions of the genome from intergenic, intronic and genic regions. Signal to background ratios were determined by first calling probes that fluoresced at intensities greater than 99% of the random probes’ signal intensities; therefore only 1% of fluorescing experimental probes should be false positives. We conducted two-color competitive hybridizations that measure differential expression from three replicates, each using RNA from independent biological extractions. Transcriptional active regions (TARs) were defined by stringing together overlapping probes showing fluorescence above a 1% false positive rate (FPR). Positive probes were joined into a TAR if they were adjacent (maxgap=0, no intermittent non-positive probe) and a TAR’s length had to be at least 45 bp (minrun=45, mid-point first positive probe to mid-point last positive probe, resulting in at least 3 adjacent positive probes for a TAR). Transcriptional active regions (TARs) were defined by stringing together overlapping probes showing fluorescence above a 1% false positive rate (FPR). The data analysis to measure differential expression of genes and of unannotated TARs was performed using the statistical software package R and Bioconductor with additions and modifications. The signal distributions across chips, samples and replicates were adjusted to be equal according to the mean fluorescence of the random probes on each array. All probes including random probes were quantile-normalized across replicates. Expression-level scores were assigned for each predicted gene based on the median log2 fluorescence over background intensity of probes falling within the exon boundaries.

Dataset Information

Measure transcript integrity using RNA-seq data

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure