Introduction
Low quality or low input RNA no longer is a showstopper for miRNA sequencing. The last iteration of our NEXTFLEX™ kit was built for these scenarios. It tolerates as little as 1 ng of total RNA (corresponding to ~50 pg of miRNA), is gel-free, and blocks adapter-dimer formation through proprietary dimer-reduction strategies. Below is a concise roadmap, written for scientists at the bench explaining best practices to squeeze the highest quality small RNA data out of sub-optimal samples.
Preserve what small RNA you have
In degraded tissue or biofluids, miRNAs survive better than mRNA because of their size, Argonaute shielding, and/or encapsulation in extracellular vesicles. However, they will fragment further if they remain warm or wet. Snap-freezing in liquid nitrogen or immediate immersion in RNAlater® (Thermo Scientific) is ideal; however, formalin-fixing of tissues is also an option. Formalin-fixed tissue should be sectioned thinly so that proteinase K and heat reversal later can reach every crosslink. Published comparisons of matched fresh-frozen and FFPE blocks consistently show high miRNA concordance even after a decade of storage, reinforcing the value of sequencing miRNAs even when mRNA integrity is lost, which is common in formalin-fixed samples.
While small RNAs are generally more resistant to freeze-thaw induced fragmentation, it is best practice to make single-use aliquots. When using limited RNA input with the NEXTFLEX™ kit, maximizing input quantity is critical. Specific amounts of RNA should be aliquoted depending on the inclusions (4 µL) or exclusion (5 µL) of T/Y RNA blockers prior to freezing and storage at -80°C. Additionally, limiting the time small RNA aliquots are stored prior to use, can have modest improvement on library preparation and ultimately data quality.
Extract the RNA, don’t enrich it yet
Small RNA libraries can be prepared using total RNA. Enriched small RNA samples can be used but is not recommended when working with challenging samples, due to loss of sample through any enrichment process.
Column-based kits that include proteinase K outperform phenol-based extraction methods for oxidised or cross-linked samples because they retain short fragments without co-precipitating salts.
It is better to co-purify total RNA rather than size-selecting at this stage as fragmented long RNAs behave like miRNAs and are easily removed later by the NEXTFLEX cleanup beads. If your yield of total RNA will be below 50 ng, add linear acrylamide or GlycoBlue™ (Thermo Fisher) during the isopropanol precipitation step to prevent pellet loss.
Measure what matters
Traditional RQS or RIN scores offer a high level view of RNA fragmentation but they are insufficient for miRNA. Instead, inspect a small RNA LabChip trace (or Agilent alternative): even a blunted 20–40 nt peak indicates that these species are present.
Quantitative reverse-transcription PCR (RT-qPCR) of a single, well-expressed miRNA in the sample (classically miR-16-5p or miR-191-5p) offers a fast test of whether an extremely degraded sample is still suitable for small-RNA-seq. Unlike total RNA metrics, the Cq value of a mature miRNA directly reports on the biochemical feature that library preparation actually needs strands with a ligatable 5′-phosphate and 3′-hydroxyl. Comparative studies show that when the same RNA extract yields Cq ≤ 30, NEXTFLEX v4 libraries routinely surpass 1 nM yield and recover >400 miRNA species, whereas Cq ≥ 33 predicts both low ligation efficiency and a collapse in complexity after PCR. This RT-qPCR assay can be used as an extraction-efficiency check: a delayed Cq together with normal spike-in recovery flags copurified inhibitors, a pattern confirmed in FFPE and plasma workflows that compared column chemistries by RT-qPCR before committing to sequencing.
Impact of RNA degradation small RNA sequencing results
When the input sample is heavily degraded, it may contain a significant proportion of 5’-phosphorylated fragments with size <16 nt, which is problematic. Those fragments will appear as libraries with a shorter size than those obtained with the positive control included in the NEXTFLEX Small RNA sequencing v4 kit. The problem of the fragments with size<16 nt is that they are notoriously difficult to map unambiguously. Bowtie, BWA and other seed-based aligners generally discard them, which will lead to a decrease on the percentage of reads mapping to small RNA.
With challenging samples, it can be beneficial to perform first a low pass small RNA sequencing run to evaluate the percentage of miRNA mapping reads observed and fine tune the amount of reads per sample that is most suitable.
To improve mapping rate in samples with excessive degradation there are two options: one is to use an upstream kit (such as the Monarch™ Spin RNA Cleanup kit (NEB) or the RNA clean & concentrator kit (Zymo) to remove RNA fragments <16 nt before library preparation. The second option is to perform polyacrylamide gel electrophoresis (PAGE) to separate the band of intact small RNA from other fragments, followed by gel band excision and purification. Please contact Revvity technical support team to discuss if this is needed and the best approach for your project.
Tuning the small RNA seq workflow
With low input or degraded samples, the effective adapter to insert ratio favours adapter dimers. For this reason, we recommend diluting the NEXTFLEX 3' Adenylated adapter and the NEXTFLEX 5' Adapter 1/4 with nuclease-free water to reduce chance of adapter dimer formation.
miRNA copy number and repertoire differ markedly among biological samples. Because of this, the manual of the NEXTFLEX Small RNA sequencing v4 kit offers guidance, but PCR cycle number must be empirically titrated so that each library attains a comparable yield while avoiding excessive duplication. When the input RNA is heavily degraded, adding one or two extra cycles over the baseline protocol offsets ligation losses and restores the library to the desired concentration window.
Small RNA sequencing depth and analysis checkpoints
Five to ten million reads per library usually recover >500 unique human miRNAs from poor-quality inputs; overshoot to twenty million if isomiR discovery is central to you, or if you have a low percentage of reads mapping miRNA.
Before mapping, trimming adapters is necessary. Aligning reads using stringent requirements in a hierarchical fashion to specifically curated individual reference databases helps ensure accurate alignments. After quality control to remove low quality reads, <16 nt reads, and reads without adapters, the resulting reads are aligned first to miRBase, remaining unaligned reads are then aligned to tRNAs and Y-RNAs, etc. Finally, if spike-ins with known relative copy amounts were included in the experiment, normalised expression can be obtained which can mitigate effects of variable sample quality. These combined analytic steps reflect best-practice bias control highlighted in recent comparative sequencing benchmarks.
You can also use specialized pipelines such as exceRpt, which employs a scoring algorithm, to evaluate the results. Both hierarchical and scoring algorithms strategies generally produce nearly identical results, each has their own pros and cons which must be weighed before use.
For research use only. Not for use in diagnostic procedures.
