Sources of non-uniform coverage in short-read RNA-Seq data

Area of Science:

• Molecular Biology
• Genomics
• Bioinformatics

Background:

• RNA splicing is fundamental to normal cellular functions and disease.
• Short-read RNA sequencing (RNA-Seq) is the standard for transcriptome quantification.
• Extreme non-uniformity in RNA-Seq coverage across transcripts is a major technical artifact, hindering accurate isoform-level analysis.

Purpose of the Study:

• To identify and investigate the sources of non-uniformity in RNA-Seq library preparation.
• To critically evaluate factors influencing RNA-Seq coverage, challenging existing hypotheses.

Main Methods:

• Exploration of eight potential sources of non-uniformity.
• Targeted experiments on fragment length, PCR ramp rate, and ribosomal depletion.
• Assessment of existing datasets with variations in sample quality, PCR cycles, and reverse transcriptase.

Main Results:

• Non-uniformity in RNA-Seq coverage is multifactorial and cannot be attributed to a single cause.
• Secondary structures unlikely to significantly interfere with reverse transcription.
• Ribosomal depletion methods do not introduce non-uniformity.
• PCR ramp rate and fragment length have minimal impact on coverage non-uniformity, contradicting previous recommendations.

Conclusions:

• Existing RNA-Seq protocols and analysis methods may need re-evaluation due to the multifactorial nature of coverage non-uniformity.
• Findings challenge common assumptions and prior publications regarding RNA-Seq artifacts.