RNA sequencing provides a measurement of the RNA molecules represented in a biological sample, but that representation can be shaped by extraction, library preparation, sequencing, and analysis. Before RNA reaches sequencing, it has already passed through several selection points, including RNA extraction. This is not necessarily a "neutral" step in the process, as RNA size, abundance, composition, sample type, and purification chemistry all influence RNA recovery. This is extremely important to note across research in miRNAs, small RNAs, circulating RNA, and low-input samples. Silica is a conventional approach to RNA purification, though silicon carbide-based purification uses a different matrix and binding chemistry that can influence RNA recovery. In this article, we will explore what factors can limit unbiased RNA recovery, and compare these two approaches. For example, in some cases showing technical bias, when small non-protein coding RNA is sequenced, the expression levels of some RNAs become artificially enhanced, and some diminished or undetectable.5 Exploring bias and how it interacts with these approaches is fundamental to understanding how to effectively approach RNA-seq with your sample.
Why RNA Extraction Matters for RNA Sequencing
When it comes to RNA extraction, your extraction quality is vital, as it can influence downstream RNA-seq and is one potential source of bias. RNA-seq is an intricate workflow of RNA extraction, library construction, sequencing, and bioinformatics analysis.6 When any of these steps within the larger process are faced with a deviation, this can complicate your results. Some of the most prominent sources of RNA-seq bias include RNA degradation, size, low-abundance transcripts, and technical steps involved in library preparation, including adapter ligation and reverse transcription.2 Extraction chemistry can affect RNA recovery, degradation, contaminant carryover, and the relative representation of RNA molecules with different properties.2 As RNA extraction is one step within a larger workflow, consistent recovery in high-quality RNA is important for reliable representation throughout the downstream processes.
Understanding RNA Binding Chemistry
In silica-based purification, RNA binds to the silica surface under defined combinations of chaotropic salts, ionic strength, pH, and alcohol. Wash steps remove proteins, salts, and other contaminants while the RNA remains bound, after which the RNA is eluted. Strong recovery of RNA is dependent on the RNA characteristics, binding chemistry, buffer conditions, and matrix properties.
Why Small RNA Is Especially Challenging
What are small RNAs?
Small RNAs are commonly defined operationally as non-coding RNAs shorter than approximately 200 nucleotides, although the precise definition depends on the study and RNA class.
- miRNAs (microRNAs)
- miRNAs are typically about 20–24 nucleotides long and regulate gene expression by binding, usually through partial complementarity, to target mRNAs. This can repress translation and/or promote mRNA destabilization and degradation.2
- siRNAs (small interfering RNAs)
- siRNAs are short RNA duplexes, commonly around 20–24 nucleotides long. One strand is loaded into an Argonaute-containing complex, which guides recognition and, in suitable cases, cleavage or repression of complementary target RNAs.
A sample may contain smaller amounts of a particular mRNA, but that molecule could still be significant. Circulating miRNAs are commonly used in studies, as they have emerged as promising, non-invasive cancer biomarkers.2 These small non-coding RNAs regulate gene expression through binding to target mRNAs, and are highly stable.2 Notably, small RNA recovery can hinge on purification chemistry, and protocol conditions.
Can RNA Extraction Introduce Sequence-Dependent Bias?
Careful selection of the RNA isolation method and consideration of the type and amount of starting material are important for avoiding biased results.3 Some conventional total-RNA protocols, including certain phenol-based and silica-column workflows, may recover small RNAs inefficiently unless the protocol and product are specifically designed to retain them. Performance depends on the chemistry, sample type, input amount, and size-selection steps.6 Because these molecules carry biological information, the extraction chemistry of the RNA can meaningfully influence which transcripts are represented in the final sequencing run.
Comparing Silica With Silicon Carbide
When comparing silica against silicon carbide, a few details stand out. Both approaches are capable of producing RNA of adequate purity and integrity, but the RNA species recovered by each matrix can differ.
This distinction is significant when the research question depends on small or low-abundance RNA. Different extraction methods are capable of recovering significantly different amounts of miRNA, even though they are all designed for RNA purification. Ahlberg et al. (2021), found significant differences in both total miRNA concentration and the recovery of specific endogenous and exogenous miRNAs from five commercial extraction kits.1 Ahlberg et al. suggested that differences in column matrix binding capacity may contribute to variable miRNA recovery.1
When directly comparing silicon carbide with silica and phenol extraction methods, Kim et al. (2015) found the silica and silicon carbide columns produced comparable RNA integrity and purity.4 However, the SiC column purified more small RNA than the silica column, and small RNA and miRNA were found to be the most abundant in the tested SiC-purified samples based on Bioanalyzer analysis.4
These findings indicate that RNA quality cannot be solely determined by the factors of concentration and purity. Two methods of extraction are capable of producing RNA with similar A260/280 ratios and integrity while differing in which RNA species are actually represented in the final preparation.
Silent Bias: The RNA You Never Measure
The most consequential extraction bias is the one that is difficult to derive in origin after sequencing. If an RNA species is poorly recovered during extraction, it may be underrepresented in downstream analysis. As a result, it may appear at lower abundance, or disappear below the detection threshold entirely.5
This becomes very important for small RNAs. Small-RNA sequencing is already vulnerable to technical biases introduced during adapter ligation, where short DNA/RNA adapter sequences are attached to the ends of the derived molecules for amplification and sequencing.5 These technical effects cause small RNAs to appear artificially enriched while others are diminished or even undetectable.4 RNA sequence characteristics, including GC content, are also factors that can influence downstream representation.5
Why Does the Difference Matter?
For low-input samples, circulating RNA, miRNA studies, and biomarker discovery, differences in RNA recovery can affect the representation of the biological insights found. El-Khoury et al. (2016) compare several commonly used RNA isolation methods across cells, plasma, and exosome samples, and found that miRNA recovery was strongly influenced by both isolation and the amount of starting material used.3 Some miRNAs were preferentially recovered by particular extraction methods, which supports the notion that a careful selection of isolation method and input material can help avoid a biased result.3 This suggests extraction chemistry influences the composition of the RNA population entering the sequencing workflow.
Why Does the Difference Matter?
In a comparative study published in the Translational Oncology journal, Norgen's Urine Total RNA Purification Kit Dx (Slurry) was compared to competitor kits, and performed strongly across key measurements in biomarker discovery.7 It ranked highest for miRNA yield, enrichment, and spike-in recovery, with the lowest PCR inhibition rate.7 The article quotes 'Norgen's method's low PCR inhibition and high miRNA yield in whole urine and supernatant fractions make it a robust choice for these sample types'.7 This study also demonstrates the notion that the purification matrix can influence which RNA molecules are recovered, and how effectively they can be detected downstream.
Summary
Ultimately, RNA extraction is an important, but overlooked source of variation in RNA-seq workflows. While downstream applications can introduce bias, the extraction method can influence which RNA molecules are recovered and carried forward into analysis. This is particularly important in preclinical applications, namely biomarker discovery and validation testing. Silica and silicon carbide purification can produce RNA of comparable quality, while differing in recovered RNA populations. Research comparing these matrices has demonstrated differences in small RNA recovery, highlighting concentration, purity, and integrity alone is not enough to fully describe the suitability of an extraction method for RNA-seq, but also by considering sample type and downstream application. Selecting a purification method that supports broad and consistent RNA recovery will help reduce one's potential source of bias before RNA even reaches library preparation and sequencing.
