How Enzyme Engineering Improves NGS Library Preparation

Next-generation sequencing (NGS) workflows depend on enzymes at every stage, yet enzyme performance is often overlooked during workflow development and execution. From DNA fragmentation and adaptor ligation to target amplification, enzyme fidelity, efficiency, and robustness can significantly influence data quality, workflow speed, and confidence in experimental results.

In this Innovation Spotlight, Emily M. LeProust, chief executive officer and co-founder of Twist Bioscience, explains how advances in enzyme engineering are improving NGS library preparation, reducing bias and sequencing artifacts, and enabling researchers to generate more accurate, reliable data from even the most challenging sample types.

Why should researchers care about enzyme engineering if their projects aren’t focused on enzymes?

Your results are only as good as the enzymes running your workflow. Most researchers have never questioned whether those enzymes are actually good enough. Enzymes aren’t passive reagents. Every time you amplify a target sequence or ligate a sequencing adaptor, the enzyme’s fidelity, speed, and tolerance for challenging conditions directly shape what ends up in your data. That matters, whether you’re profiling tumor mutations from formalin-fixed paraffin-embedded samples, chasing a low-frequency variant in cell-free DNA, or trying to extract signal from a limited biopsy.

A polymerase with even modest amplification bias can introduce artifacts that look like real variants. A ligase that struggles under suboptimal conditions loses fragments you can’t recover. Advances in enzyme engineering don’t ask you to change your research focus. They deliver measurably better performance from the tools you already depend on: more uniform coverage, fewer amplification artifacts, and faster workflows so that the limiting factor in your research is the biology, not the library prep.

How can enzyme engineering specifically improve NGS data?

NGS workflows often rely on multiple enzymes to produce a sequencing-ready DNA library. Samples need to be fragmented, which can be done mechanically or enzymatically. Then sequencing adaptors must be ligated to DNA fragments, enabling them to be read by the sequencer. Finally, you may want to enrich and amplify specific target sequences, which requires the use of a polymerase.

At each step, enzyme performance can have a significant influence on the efficiency and accuracy of your workflow. Enzymatic fragmentation has many advantages over mechanical methods, but if the conditions aren’t optimized, fragment lengths can vary considerably. This can result in greater fragment overlap, oversequencing in some regions, and overall reduced economy of sequencing.

Amplification steps are also at the mercy of enzyme performance. Polymerases can be biased for specific sequence traits or show reduced fidelity in complex regions. This can lead to nonuniform amplification of target sequences, artificial mutations, and dropouts. Similarly, ligases responsible for attaching sequencing adaptors to DNA fragments can show biases that slow or block the ligation process, leading to lost fragments.

There are of course ways for you to overcome these limitations, but these usually come with a steep cost, in dollars, time, and confidence. Alternatively, enzyme engineering could be used to build better enzymes that reduce bias and improve the accuracy of sequencing readouts.

Do researchers have to engineer their own enzymes to realize these benefits?

Fortunately, no. Enzyme engineering is no small feat, and most laboratories won’t have the resources or expertise needed to optimize enzymes across multiple competing properties.

That’s where Twist’s position is unique. We have built a DNA synthesis platform that enables us to synthesize virtually any DNA sequence at scale. When combined with artificial intelligence (AI), we are able to design and synthesize diverse libraries of enzyme variants for wet-lab screening. Because we are both a DNA synthesis platform and an NGS solutions provider, we sit at a rare intersection: we can design and synthesize large libraries of enzyme variants and validate them directly against the sequencing workflows our researchers use every day. That closed loop of design, synthesize, test, and iterate is what makes it possible to engineer enzymes that perform under real experimental conditions, not just idealized assay settings.

We now offer library preparation kits that are fortified with both optimized solutions and engineered enzymes, each of which is carefully designed to enable more efficient and reliable production of NGS data, even from some of the most challenging sample types.

Which enzymes have Twist engineered to improve NGS data quality?

Two enzymes are at the core of what we’ve built: a novel polymerase and a re-engineered T4 DNA ligase. Together, they address the two most consequential failure points in library preparation.

For amplification, there are many features in a polymerase that need to be balanced. Speed, for example, often comes at the expense of accuracy or increasing bias. Through an iterative design, build, test, learn cycle, we homed in on the Twist TrueAmp polymerase, which possesses an optimized balance that prioritizes accuracy and uniformity. The enzyme has a specialized proofreading domain that monitors incorporation accuracy and helps to reduce artificial mutations, even across long homopolymer stretches. It is also capable of reducing C-to-T transition mutations, which are common artifacts in formalin-fixed samples. Not only does this help you to improve sequencing accuracy, but it also preserves your library’s complexity.

Additionally, we engineered a high performance T4 DNA ligase which can more efficiently convert A-tailed dsDNA substrate into an amplifiable double-ligated product. This enzyme can produce higher yields in shorter amounts of time, resulting in more efficient and uniform libraries. Crucially, the enzyme is also engineered to be more tolerant of pH and salinity variations in the reaction solution. This enables library preparation to proceed with fewer purification steps, ultimately enabling much faster workflows.

Together with an optimized enzymatic fragmentation solution, these enzymes greatly improve the library preparation process by improving amplification accuracy, even across complex sequences; increasing library complexity and uniformity; and enabling faster, more efficient workflows.

Is there anything else that you would like to share with our readers about enzyme engineering and its application in NGS?

I’ll just emphasize that you don’t have to settle for suboptimal workflows simply because they’re familiar. Science is advancing on many fronts, some of which can help you become more efficient and more accurate with your studies. At Twist, we’re always looking for ways to bring these innovations to you, whether that’s engineering enzymes for better library preparation kits, or white labeling custom engineered enzymes for your specific NGS needs. If you want better than the norm, talk to us.