Overview of Strand-Specific RNA-Seq Library

In the era of genomics, deciphering the intricacies of gene expression is paramount to understanding cellular processes, disease mechanisms, and biological diversity. RNA sequencing (RNA-Seq) has revolutionized transcriptomics by enabling comprehensive profiling of RNA molecules within a sample. To further enhance the accuracy and depth of transcriptomic analysis, researchers have turned to strand-specific RNA-Seq libraries. These libraries retain crucial orientation information, empowering researchers to unravel the complexity of transcriptional regulation, identify natural antisense long non-coding RNAs (lncRNAs), and probe gene structure and function.

The Significance of Strand-Specific RNA-Seq Libraries

Conventional RNA-Seq libraries lack information about the strand from which a transcript originated, limiting accurate gene quantification and the detection of variable splicing events. Strand-specific libraries address this limitation by preserving the orientation of RNA molecules, allowing researchers to discern whether reads are derived from the positive or negative strand. This additional dimension of data provides a richer understanding of gene expression dynamics, facilitates the investigation of antisense transcripts, and enhances the analysis of gene structure and function.

Methods for Strand-Specific Library Construction

Several methods have been developed to construct strand-specific RNA-Seq libraries, which can be broadly classified into two categories: junction-based methods and chemical modification-based methods.

• Junction-based methods: These methods involve adding different junctions at the 3' and 5' ends of the RNA molecules to label the strand orientation. There are several techniques based on this principle, such as:

a. dUTP method: In this method, the second strand of cDNA synthesis includes the incorporation of dUTP instead of dTTP. After cDNA synthesis, the dUTP-containing strand can be selectively degraded using uracil DNA glycosylase (UDG) treatment. This allows the remaining strand to be used for library preparation and retains the strand information.

b. Ribo-Zero method: This method involves using specific ribonucleases to selectively degrade RNA molecules from one strand while preserving the RNA from the opposite strand. The remaining RNA is used for library construction.

c. Template-switching methods: Techniques such as SMART (Switching Mechanism at the 5' end of RNA Template) utilize template-switching oligonucleotides (TSO) during cDNA synthesis. TSO contains a specific sequence that facilitates the addition of a unique primer-binding site, which enables strand-specific library construction.

Methods for strand-specific RNA-Seq. (Levin et al. 2010)

• Chemical modification-based methods: These methods utilize chemical modifications to specifically label one strand of the RNA molecule. The modifications can involve base modifications or degradation of one strand. Some examples include:

a. Bisulfite treatment: Bisulfite treatment converts cytosines to uracils, and the treatment is strand-specific. By sequencing the treated RNA, it is possible to determine the original strand information.

b. RNase H treatment: RNase H specifically degrades the RNA strand of RNA-DNA hybrids. By selectively degrading one strand, the remaining DNA strand can be used for library construction.

Suitability of Different Methods for RNA Samples

• Small RNA samples: Small RNA molecules such as miRNA, piRNA, tRF&tiRNA, tRNA, snRNA, and snoRNA are typically shorter in length and lack poly(A) tails. For constructing strand-specific libraries from small RNA samples, RNA ligation-based methods that join junctions at the 3' and 5' ends of the RNAs are commonly used.
• mRNA/LncRNA/CircRNA, etc.: These longer RNA molecules can be synthesized using random primers or oligo dT primers for cDNA synthesis. For strand-specific library construction, methods like the dUTP method or template-switching methods can be employed.
• Difficult-to-amplify or low-input samples: In cases where the RNA yield is low, such as with small cell populations or challenging sample types like serum/plasma or exosomes, specialized techniques like the SMART technique (with splice sequence addition) can be used to synthesize the first strand of cDNA and retain strand origin information.

Comparison of Various Strand-Specific Libraries

When comparing the sequencing results of various strand-specific libraries, several factors come into play.

• Strand Specificity
  Strand specificity refers to the ability of the library construction method to accurately determine the origin of the RNA transcript strand, whether it is derived from the positive or negative strand. Methods such as dUTP two-stranded labeling, Ribo-Zero, and template-switching have demonstrated high strand specificity. However, it is important to note that the degree of strand specificity can vary among different library construction methods.
• Library Complexity
  Library complexity refers to the diversity and richness of the RNA molecules captured within the library. It is influenced by factors such as the efficiency of cDNA synthesis, fragmentation methods, and bias introduced during library preparation. The impact of library construction methods on library complexity can vary, with some methods exhibiting higher complexity compared to others.

Key criteria for evaluation of strand-specific RNAseq libraries. (Levin et al. 2010)

• Mapping Efficiency
  Mapping efficiency relates to the proportion of reads that can be successfully aligned to the reference genome or transcriptome. Methods that generate high-quality libraries with low levels of adapter contamination, PCR duplicates, or sequencing errors tend to exhibit better mapping efficiency.
• Coverage Continuity
  Coverage continuity refers to the evenness of read distribution along the entire length of the transcript. Methods that result in uniform coverage across the transcript are desirable as they enable accurate measurement of expression levels and detection of alternative splicing events. In contrast, methods that introduce biases or uneven coverage may lead to incomplete representation of the transcriptome.
• Consistency with Known Annotations
  Comparing the sequencing results with known annotations, such as annotated gene models or transcript databases, allows assessment of the accuracy and reliability of the library construction methods. Consistency with known annotations suggests that the method is capturing the expected transcriptome features accurately.
• Accuracy of Expression Profiles
  Accurate quantification of gene expression is a key goal of RNA-Seq. Library construction methods that provide consistent and reliable expression profiles, both in terms of relative expression levels and absolute quantification, are highly valuable for downstream analysis.

dUTP Second Strand Marking Method in Yeast Transcriptome

A comparative analysis of various strand-specific library construction methods using the Saccharomyces cerevisiae transcriptome as a benchmark revealed significant differences in sequencing results. The evaluation encompassed parameters such as strand specificity, library complexity, homogeneity and coverage continuity, consistency with known annotations, and accuracy of expression profiles.

Comparative analysis of strand-specific RNAseq libraries. (Levin et al. 2010)

Among the tested methods, the dUTP second strand marking method emerged as the top performer based on its superior performance and simplicity. This method outshined the others in terms of several key metrics. Firstly, it exhibited a higher proportion of uniquely mapped reads in both single-end and double-end sequencing data, indicating enhanced mapping accuracy and reduced ambiguity. Accurate mapping is crucial for reliable gene expression quantification and identification of variable splicing events.

Furthermore, the dUTP method demonstrated a more balanced distribution of transcript coverage across the 5' and 3' ends, ensuring comprehensive coverage of the entire transcript. This is crucial for capturing the full complexity of gene expression patterns and accurately characterizing transcriptional landscapes. In contrast, other methods may exhibit biases toward either the 5' or 3' ends, leading to incomplete coverage and potential distortion of expression profiles.

Importantly, the dUTP library exhibited high inter-sample reproducibility, suggesting consistent and reliable results across multiple experiments. Reproducibility is a critical aspect of any sequencing method as it ensures the robustness and consistency of findings, allowing for meaningful comparisons between samples and datasets.

Reference:

1. Levin, Joshua Z., et al. "Comprehensive comparative analysis of strand-specific RNA sequencing methods." Nature methods 7.9 (2010): 709-715.