February 1, 2015 (Vol. 35, No. 3)
Adam R. Morris, Ph.D. senior scientist Bioo Scientific
Masoud M. Toloue, Ph.D. vice president Bioo Scientific
Randomized Adapter Strategy for Library Preparation Reduces Ligation Bias and Increases Accuracy of Small RNA-Seq
The study of small RNAs, including miRNAs, siRNAs, and pi-RNAs, is an ideal application of next-generation sequencing (NGS) technology. Although methods such as quantitative PCR and microarray analysis are useful for relative quantification of small RNAs, they suffer from two major drawbacks.
The first is that these methods are hybridization-based, which presents problems when trying to discriminate two small RNAs whose sequence differs by only a nucleotide or two. The second drawback is that both of these methods are only able to interrogate an a priori determined set of small RNAs, which both limits the scope of studies and prevents discovery of new small RNAs.
Both of these drawbacks of hybridization-based methods are addressed by using NGS for small RNA studies, as NGS can reliably distinguish small RNAs that differ by only a single base, and NGS is not limited to the study of a predetermined set of sequences. However, a major drawback of NGS methods for the study of small RNAs is the substantial bias that has been shown to exist in traditional library preparation protocols. This bias has been shown to be introduced during the two ligation steps, and the combined effect of the bias introduced in these steps results in some small RNAs being ligated to adapters much more efficiently than others.
Some of the studies that demonstrated the substantial bias introduced by RNA ligases showed that this bias resulted from the adapter sequence proximal to the ligation junction, and that adapters with 2–4 randomized bases at this junction could be used to substantially reduce ligation bias.
Jayaprakash et al. first demonstrated that NGS libraries prepared with this strategy showed little evidence of ligase bias and that data generated from these libraries correlated well with microarray and qPCR data.1 Bioo Scientific has obtained an exclusive license on this patent pending technology and has since developed a library preparation kit for high-throughput sequencing of small RNAs.
Bioo Scientific has combined the power of randomized adapters with a streamlined and user-friendly protocol in the NEXTflex™ Small RNA Sequencing Kit v2. Both the 3´and 5´ adapters included with this kit contain four (bias-reducing) randomized bases at the ligation junction, and the workflow has been optimized to be more streamlined and allow the use of various sample types (Figure 1).
The kit makes use of magnetic bead-based cleanups in the place of column-based cleanups and ethanol precipitation used in other protocols, resulting in a more user-friendly workflow and less risk of losing some or all of a sample. The kit also contains all necessary enzymes, requiring the user to provide only AmpureXP and common lab materials.
Other published small RNA-Seq protocols incorporating randomized adapters all require the cutting and eluting of the ligation products of the desired size from a denaturing TBE-PAGE gel, which is a long and tedious process that often requires an overnight elution and an ethanol precipitation, to deplete excess 3´ adapter following ligation. In contrast, the NEXTflex Small RNA Sequencing Kit v2 uses AmpureXP beads, isopropanol, and a proprietary Adapter Depletion Solution to deplete excess 3´ adapter in under an hour, significantly simplifying the procedure.
Demonstration of Reduced Bias
In order to demonstrate the reduction in bias achieved by the NEXTflex Small RNA Sequencing Kit v2, a “miRNA calibrator” sample of 24 synthetic miRNAs mixed in equimolar amounts was created. These miRNAs were chosen to represent a variety of sequence combinations at the 3´ and 5´ ends. 5 ng of this calibrator sample was used to prepare libraries in triplicate with the NEXTflex Small RNA Sequencing Kit v2, which uses adapters with randomized ends, and a traditional small RNA-Seq protocol, which uses standard (nonrandomized) adapters.
These libraries were then sequenced together on an Illumina MiSeq, and the proportion of reads mapping to each miRNA present in the calibrator libraries was determined. The triplicate values were then averaged and plotted as a pie graph to demonstrate the relative proportion of reads that aligned to each small RNA in the sample (Figure 2). This analysis clearly demonstrates the substantial reduction in bias when using the latest improvements in small RNA-Seq technology.
In order to demonstrate the ability to prepare small RNA libraries from various sample types with the NEXTflex Small RNA Sequencing Kit v2, libraries were prepared from total RNA from a variety of human tissues: brain, colon, lung, thymus, esophagus, and spleen. One microgram of total RNA starting material was sufficient to prepare libraries from all tissue types tested. Between 42% and 58% of reads in these libraries mapped to human miRNAs found in the miRBase database.
While much effort is being put into understanding microRNA expression and the importance these small RNAs play in gene regulation, research is limited by skewed expression representations introduced during small RNA library preparation. NGS is an ideal technology for measuring small RNA expression, however, results from deep sequencing, microarrays, and qPCR often do not agree, making it difficult to extract conclusions. By combining adapters with randomized ends and a streamlined, user-friendly protocol, the NEXTflex Small RNA-Seq v2 kit allows for the creation of small RNA libraries with reduced bias from a variety of sample types. The reduced bias achieved by this kit will allow researchers studying small RNAs to make discoveries that may have been overlooked otherwise and allow for a greater understanding of many aspects of small RNA biology.
Adam R. Morris, Ph.D., is senior scientist and Masoud M. Toloue, Ph.D. ([email protected]), is vp, genomic research at Bioo Scientific.
1. Jayaprakash, A.D., et al., Identification and remediation of biases in the activity of RNA ligases in small-RNA deep sequencing. Nucleic Acids Res, 2011. 39(21): p. e141.