Breaking Through Bias to Redefine Precision in Small RNA Library Prep

Deyra Rodriguez:
Hi, everyone. Welcome to our webinar, Breaking Through Bias to Redefine Precision in Small RNA Library Prep. My name is Deyra Rodriguez. I'm a Product Marketing Manager and together with Heather Raimer Young, Development Scientist, we'll be introducing our newest solution for small RNA library prep.

This is our agenda for today. I'll start by reviewing some background on small RNAs and what we consider bias in small RNA library prep. Then I'll share the key features of our new product. At this point, I'll transfer the floor to Heather who will walk you through the method and share data showing the product's performance, and we will conclude with our question and answer session.
Let's get started by talking about small RNAs. Small RNAs represent a diverse class of RNA. Many types have been identified and new ones continue to be discovered. These RNAs are small, typically less than 200 nucleotides in length. They're usually non-coding, meaning that they don't code for proteins, but instead they play a key role in regulating gene expression through different mechanisms.

In this diagram, you can see some of the most studied types of small RNA and their mechanism of action. For example, some small RNAs work to silence other RNAs. In this category, we have microRNAs and piRNAs. MicroRNAs bind to complementary sequences in mRNA and this binding causes that mRNA to be degraded or inhibits its ability to get translated. PiRNAs or PIWI-interacting RNAs are mostly active in reproductive cells. Here, they guide PIWI proteins to cleave potentially harmful RNA such as transposon RNA.

Other small RNAs are involving RNA splicing. This is a case of snRNAs or small nuclear RNAs, which are core components of the spliceosome. tRNAs are involved in translation. They transport amino acids to the ribosome for protein synthesis. And finally, snRNAs or small nuclear RNAs, which guide chemical modifications such as methylation and pseudo relation of other RNAs, including ribosomal RNA and tRNA.

Why are small RNAs important? Well, small RNAs are key for maintaining proper cellular function and health. By regulating gene activity, small RNAs help orchestrate complex processes like development and cell differentiation, stress responses, and immune defense. They also play a role in maintaining the epigenetic landscape, helping preserve a cell's identity this way. But small RNAs are not just relevant for normal cellular functions. They also change their expression levels in response to diseases. This has been observed in conditions including cancer, metabolic disorders, neurological and infectious diseases. Because of this disease associated change into expression level, small RNAs can serve as clinical biomarkers helping in both the diagnosis and prognosis of diseases.

It's worth noting that small RNAs have been a significant topic in research for the past 25 years. A quick search in PubMed shows a positive trend in the number of publications mentioning small RNA. And just last year, the Nobel Prize in physiology or medicine was awarded to Victor Ambros and Gary Brufcon for the discovery of microRNAs and the role in post-transcriptional gene regulation. It's an exciting time to be in the small RNA field.

Sequencing is the preferred method for discovering and studying small RNAs. However, when preparing a small RNA library for sequencing, bias can be introduced. Addressing this bias is a challenge in library preparation. So what do we mean when we say bias? When we talk about bias in small RNA library prep, we're referring to the inaccurate representation of the different small RNA molecules in a sample.

Imagine you have a sample of total RNA and you're interested in sequencing its small RNAs. Typically, around 5% of that sample is the small RNA population you want to capture. Now, if you use a small RNA library prep kit that is biased, when you analyze your data, the results will likely not represent the small RNA population in the original sample. In the diagram, you can see some RNA species missing, that green one in the center, and others represent it at different levels compared to the starting material. This could lead to misinterpretation of biological data, and of course this is not what we want.

On the other hand, when you use a low-bias library prep kit, your data will better represent the population in the sample, helping you understand the biological function more clearly. This is what we want. We want to capture all the small RNA species present and do it accurately.

But where is bias coming from? Bias can be introduced at different steps in the library prep process. The typical ligation-based small RNA library prep workflow has several key steps. An initial adapter ligation step followed by reverse transcription, PCR amplification, and gel site selection. Bias can be introduced at any of the steps. However, research shows that the primary source of bias is that adapter ligation step.

The efficiency of adapter ligation can be influenced by different factors, including the sequence preferences of the ligases used, the secondary structures of both the small RNAs and the adapter used for library preparation, and modifications on small RNAs such as 2-prime O-methylation. At NEB, our scientists have been working hard to understand and mitigate the sources of bias.

We launched our first small RNA library prep kit in 2011. Since then, we have continued to study and improve the library prep process. Our focus has been the challenge our customers face with one major challenge, of course, being bias. This effort is reflected in two key publications from our research group. The first publication by Ryan Fuchs and colleagues delves into understanding the sources of bias, showing that the secondary structures of adapters and RNA are key contributors.

The second publication led by Sean Maguire introduces a new method that uses randomized splint ligation to increase sensitivity and dramatically reduce bias. This is the method that our product development team optimized, streamlined, and turned into a robust user-friendly solution.

The result is the NEBNext Low-bias Small RNA Library Prep Kit. We're excited to present this kit, which stands out for having the lowest bias among commercially available kits. It accommodates a wide range of input amounts. You can use total RNA inputs ranging from as high as 1 microgram to as low as 0.5 nanograms. When using enriched RNA, the kit supports input from 5 nanograms down to 50 picograms of small RNA.

Another key feature of this kit is its fast gel-free workflow that can be completed in approximately 3.5 hours. It's compatible with 480 unique dual indices for multiplexing. These are available for purchase separately. Additionally, no protocol changes are needed to capture those typically challenge in 2-prime O-methylated species such as piRNAs and microRNAs implants. Finally, the kit has an 18-month shelf life, which makes it convenient for ordering and storage.

Now, I would like to pass the spotlight to Heather who will delve deeper into the method of our new kit and its performance.

Heather Raimer Young:
Thank you. As Deyra mentioned, my name's Heather Raimer Young. I'm a development scientist here at NEB working on our next gen sequencing group. I'm excited to share with you the method and product performance data for our new Low-bias Small RNA Library Prep Kit.

I'll start with the molecular overview of the method. You start with a RNA input that has the five-prime monophosphate and a three-prime hydroxyl group. This then goes into three-prime adapter ligation where we add in our double-stranded splint adapter to the three-prime end. The splint being that lighter colored single-stranded overhang. And this is one means by which we reduce bias with this kit.

Following that step, we go through an excess adapter removal where we should remove any residual unligated adapter. This is then followed by the five-prime adapter being ligated. Again, this is a double-stranded splinted adapter design. Additionally, in this step, we're creating an RT primer out of the bottom strand of the three-prime adapter. This allows us to go immediately into reverse transcription where we prime off of the bottom strand of the three-prime adapter, reverse transcribe creating our cDNA, and again, because of the way we've designed this, the randomized splints on either adapter are not incorporated into the final library.

This goes into then our bead size selection where you're removing any high molecular weight material that's not of interest and any residual adapter from the five-prime ligation step. You then go into PCR amplification where we're using unique dual indexing primers, which allow for a much higher level of multiplexability with this low-bias small RNA kit. Following PCR, you go through bead purification and come out with your final library ready for quantification and sequencing.

Every step of this workflow has been optimized. We've shortened reaction times and moved to bead-based size selection and final library purification, which can be used with any input to create this streamlined workflow that can be completed in approximately three and a half hours.

Now moving into our performance data, we'll start with the assessment of bias. We compared the NEBNext Low-bias Small RNA Library Prep Kit to Revvity's NEXTFLEX Small RNA Sequencing Kit B4, Qiagen's QIAseq microRNA Library Kit, and the Illumina TruSeq Small RNA Library Prep Kit. In this, we used a set of synthetic control microRNA-like sequences that are an equimolar mix of 100 sequences that we have here at NEB. We input 300 piece grams of this RNA as it's an enriched small RNA sample. We sequenced on an Illumina instrument to 5 million reads per library, performed adapter trimming and aligned to these synthetic controls.

Because we used an equimolar mix, we expect that each sequence should be equally represented in the final sequencing data. So we're able to compare how many reads we observe to how many are expected to assess how well we represent the true mixture. So a perfect observed-to-expected ratio is shown here by the value at 100 indicated by the dark line. We assess what is seen within twofold of that idealized ratio and that's shown by the shaded region.

The NEB kit shown in green shows most of the data tightly packed within that twofold region with 90% of the points falling within. And that's consistent across three replicates indicating a very tight distribution, very low bias. That's compared to Revvity, Qiagen and Illumina where they have from the mid-teens to mid 30s of the sequences within twofold and many more falling underrepresented or overrepresented, indicating that this NEBNext Low-bias Kit does in fact produce libraries with the lowest bias.

So what does having the lowest bias allow us to do? We tested the same four kits using total RNA from human brain using inputs from 1 microgram down to 500 picograms while only tested within the recommended ranges for each individual kit. We sequenced libraries on the Illumina NovaSeq 6000 with 15 million reads per library. And then the analysis followed adapter trimming, alignment to the human genome and using a star reference supplemented with additional tRNA, rRNA and piRNA sequences.

First, we'll look at final library yields. NEB in green is shown from high input to low input where we can see a robust yield across the entire input range that stays fairly consistent, and when compared to the competitors, is similar or greater than their yields. But when I draw your attention to the line of PCR cycles, you'll be able to see that NEB's kit uses two to four fewer cycles for the same inputs compared to our competitors. So similar robust yields, fewer PCR cycles.

Next, we look at unique microRNA detection. Shown on the top are cumulative histograms where the count of unique microRNAs detected is on the vertical axis. Supporting reads is on the horizontal axis, and we're divided out by inputs from high to low across the graph. NEB is shown in green. And for each input, NEB has higher or comparable microRNA across the whole range of supporting reads. If we then compare for each input at 10 reads of support, that's what's quantified in the bottom graph where you can see NEB having higher or comparable microRNA detection across all inputs and still the same microRNA detection, even with 500 picograms of input, which were the only kit to go down to.

Next, we want to highlight the ability of the NEBNext Low-bias Small RNA Kit to detect two-prime O-methylated small RNAs. To do this, we use total RNAs from human testes where piRNAs are abundant and have two-prime O-methylations on their three-prime end, along with a series of plants, including Arabidopsis, rice, and soy, which have two-prime O-methylation of their microRNAs. We tested this between NEBNext, Revvity and Qiagen. Each RNA used 100 nanograms of input. We sequenced libraries to 15 million reads a piece on a NovaSeq 6000. Their analysis was similar to our previous study.

Starting with the detection of piRNAs from the human testes, again, we have a cumulative histogram this time showing unique piRNA counts in the 1000s on the vertical axis compared to the supporting reads for these observations. Again, NEB in green has a much higher and more robust detection of piRNAs compared to either of the competitors. And this is all without having to modify the protocol for the detection of these piRNAs.

This observation holds when we look at two-prime O-methylene microRNAs in plants. Across Arabidopsis, rice and soy, these cumulative histograms again show that NEB shown in green detects more microRNAs in each of these plant species compared to our competitors.

Finally, we'd like to showcase the ability of the NEBNext Low-bias Small RNA Kit to work with more degraded, clinically relevant samples such as FFPE RNAs. So to do this, we use total RNA isolated from either FFPE samples of human breast tissue or human lung. Each of these FFPE RNAs had low RIN values of 2.4 and 2.6 respectively, and very low DV200s as indicated on the slide. We used for this 100, 10 and 1 nanograms of input sequenced again on a NovaSeq 6000 to 15 million reads in similar analysis as our previous slides.

Here, I'm showing the data for FFPE breast samples on the left and FFPE lung samples on the right. In the top left of each are the final library yields where we see across the inputs tested, we have robust yields even with the same PCR cycles used for the high quality brain samples used in an earlier study. To the right, we have the library profiles from TapeStation D1000 data overlapping each of the inputs, showing that across inputs we get comparable library profiles in a very tight library profile as expected for small RNA libraries, even with these highly degraded samples.

Then in the bottom, we're showing replicate-to-replicate and input-to-input comparisons from microRNA detection. On the top left to bottom right diagonal of each is the replicate-to-replicate comparisons where we have very strong correlations across replicates. Then in the input-to-input comparisons, I'll draw your attention to the top right of each section, which compares 100 nanograms to 1 nanogram input for each sample. We see that in the correlations, you can find that even a low input degraded sample has a similar quality of data to a high input, which gives you high confidence to go in with your low input degraded samples to this low-bias small RNA kit.

Overall, with the performance data I shared with you today, I hope you've been able to appreciate the streamlined workflow of the NEB Next Low-bias Small RNA Library Prep Kit. We maintain the lowest bias. We have robust yields across a broad input range while using fewer PCR cycles. We can consistently detect a high number of microRNA across this whole input range. We identify more two-prime O-methylated small RNAs without the need to modify the protocol. And we're compatible with FFPE RNA samples, even those that are highly degraded or need low inputs.
And now, I'd like to pass things back to Deyra so she can close this out.

Deyra Rodriguez:
Thank you, Heather. And thank you all for joining us today to learn about the NEBNext Low-bias Small RNA Library Prep Kit. We hope you found the webinar informative and that you are as excited as we are about how this kit can support and strengthen your research. With that, let's open the floor to your questions.

Thanks for your questions. We're seeing them come in. Okay. We'll address the first question here. "When will the kit launch?"

Well, that is a good question. We are gearing to launch our kit in the next few days, so stay tuned. If you have registered for this webinar, you will receive an email notifying you when this occurs. Also, feel free to reach out to your account manager for details and information. Also, you will have the opportunity to request a sample for this kit if that's something you would like to do. A popup or a window have shown throughout the webinar. So if that's something you're interested in, please fill out that information. So we'll have that available to you soon in the upcoming days.

Heather Raimer Young:
All right. And our next question asked about that, "This is based on the Maguire 2020 NAR paper, and have we taken into account the effect that secondary structure has on the second ligation step as this wasn't mentioned in that paper? Can you comment on the sequences that aren't within twofold of expected recounts?"

So, thanks. That's a great question. So in this protocol, we do have an initial denaturing step of the RNA before we go into our first adapter ligation. Additionally, both of the splint adapters for the three-prime and the five-prime have those randomized splint sequences, which helped with the capture and with these structures. And when it comes to what fell without the twofold range? Since there were so few sequences outside of that range with the NEB kit, when we looked into it, we couldn't define any specific sequence characteristics.

Deyra Rodriguez:
All right. Next question. "Are the synthetic microRNAs available for purchase? Can I get those?"

Well, those are not currently available for purchase, but we're always looking for new product ideas. So stay tuned. That's a good idea.

Heather Raimer Young:
So next, I had a quick question on, "What's the overhang sequence of the splint oligos?"

For both of the adapters, these are randomized bases. So completely random bases there. But the important thing is the way that we treat the libraries and through the process, these sequences don't make it into the final library sequences. So you don't have to be concerned with additional trimming in your analysis. And we provide the necessary sequences for trimming the adapters with the protocol.

Deyra Rodriguez:
All right. Next question, "Why do you see increased library yield at 10 nanograms input compared to 100 nanograms?"

That is due to a function of the different PCR cycles.

Heather Raimer Young:
So next question we have is, "Does the Low-bias Small RNA Library Prep Kit also capture circular RNAs?"

So the important thing with any RNA that we're trying to capture this is that they have those five-prime monophosphate and three-prime hydroxyl ends. That's necessary for the ligation chemistry. So it's not something we've tested specifically to look at circular RNAs, but NEB is always working on solutions for these various types of samples.

Deyra Rodriguez:
All right. Following question, "Does your library support UMIs?"

Well, at this moment, the kit is compatible with UDIs, but UMIs is something that we are always interested. It's definitely an REF interest that we are exploring.

Heather Raimer Young:
All right. Next question that came in, "Did you try sequencing RNA from exosomes with this kit?"

So internally, we have not tried that, but we have had early access testers of the kit make libraries and sequence from exosomes and they've had good success.

Deyra Rodriguez:
Okay. Let's see the next one, "I'm familiar with the Sean Maguire publication. In what does it differ from this kit?"

Well, for the development of the kit, we really have explored every single step and visited every single step in the protocol. Some of the key differences is the time. We made it a lot faster. Very importantly, the kit is compatible with UDI indexing primers, where in the publication, that is not the case. We have expanded the info range that is shown in the publication and definitely optimized and combined some of the enzyme steps. So same principle, different optimization for the kits.

Heather Raimer Young:
Alrighty. And the next question is, "How is the performance of RNA with low RIN values?"

So thank you for that. We did show in the webinar, the performance with two different FSPE RNAs. Those had RIN values down at 2.4 and 2.6, also with low DV200 scores. So with that, we're able to get good quality libraries. Obviously, every sample is going to be different. And so, there may be some minor optimizations that may be needed for the user in their specific sample, but we've had good success even with low quality RNA.

Deyra Rodriguez:
All right. Next question, "What formats will the kit be in?" We will have it available in 96 reaction and 24 reaction format.

Heather Raimer Young:
Alrighty. Another good question here, "How do you ensure that adapter dimers don't overwhelm the bead-based dye selection?"

Thanks for that. One important step is after that initial three-prime adapter ligation, we do have an enzymatic excess adapter removal step. So that does a lot of the work to make sure there isn't any residual amounts of the three-prime adapter there to be compatible for ligation with the five-prime adapter. So a lot of that is just the workflow itself. So thanks for asking.

Deyra Rodriguez:
Let's see. One more question, "Do you recommend an RNA extraction kit?"

You really can use the RNA extraction kit of your choice. Just make sure that the kit that you're using captures those small RNA samples. In-house, we use our Monarch Spin RNA isolation kit, which is compatible for this application. So that's one option for you.

Heather Raimer Young:
All right. Another good question we got here is, "What is the range of RNA size that can be analyzed by the kit?"

That's a great question. So we're focused primarily on small RNAs. And with the size selection recommendations we have in the kit, we say that up to approximately 120 nucleotide RNAs will be in there. And at the lower end, we've primarily focused on microRNAs and that's the shortest RNA we have specifically looked at. If there's shorter RNAs of interest to anyone out there, they may be captured by the kit, but we have not specifically had samples to look at those sizes.

Deyra Rodriguez:
All right. Another question, "Does 480 UDI mean that we can multiplex 480 samples?"

That is correct. We have 480 unique dual indices. So if you have the sequencing capacity, that's something that you can definitely do.

Heather Raimer Young:
Alrighty. Another question we have here is, "Can the kit be used for ribosome profiling?"

So that's something we're very interested in and some early results from our beta tester says that yes, this would be successful with ribosome profiling.

Speaker 3:
We are coming up on time. So we're going to have time for one more question.

Deyra Rodriguez:
All right. "Is this kit automation-friendly?"

Yes, we have developed a protocol with automation in mind. With the bead size selection, you don't have to do a gel. And the volumes were all developed in mind for automation. So yes, it is automation-friendly.

Speaker 3:
Thank you so much, Deyra and Heather, and thank you all for joining us today. If we didn't have a chance to get to your question here in the live portion of our event, we will be sure to follow up with you online afterwards.

The webinar is available on-demand as well after today, so feel free to share with your colleagues. And again, if you are interested in a sample, feel free to fill out the survey that you see in our webinar console. If you have any additional questions, feel free to email us at Webinars@NEB.com. Thank you so much and enjoy your day.