Guide to NGS Sample Preparation

Next generation sequencing (NGS) sample preparation can be time consuming and perhaps even frustrating, especially when working with damaged or degraded samples. NGS targets can come from tissue or cell lines, blood or semen, soil or sarcophagi. They can come directly from the source or be converted from RNA to cDNA. Ultimately, though, the input for a sequencer—whatever the platform—is stretches of target DNA, of an appropriate length, ligated to appropriate adapters, at an appropriate concentration. Damaged samples, such as from formalin-fixed paraffin-embedded (FFPE) clinical biopsies, environmental samples exposed to the elements, and archeological finds, are more likely to have damage-induced mutations and yield erroneous sequences, if they can be sequenced at all.

Here we offer some tips and hints—some obvious, others less so—on how to get those samples (difficult or not so much) prepped and ready to sequence.

For a start

Any sample prep should start with a clean lab area and equipment, with special precautions to prevent RNA degradation, says Susan Magdaleno, Director R&D at Thermo Fisher Scientific. Tissues tend to have more RNases, so “it’s especially important to homogenize quickly in the lysis buffer.” As for cultured cells, they can be scraped or enzymatically released, which will not interfere with nucleic acid extraction. Researchers can also preserve RNA integrity by storing cells or tissue samples in RNALater or similar stabilization medium.

Mundane as it may sound, understanding the containers samples come in, and using the right ones, can help avoid pitfalls. For example, de-paraffinization of FFPE tissue “can be simplified using sample tubes that help minimize that variability that can come from these steps,” notes Kelli Bramlett, Senior Director R&D at Thermo Fisher Scientific. Similarly, it’s important to know what type of tube blood was collected in—with or without preservative or EDTA, for example—and follow the appropriate recommendations to obtain peripheral blood mononuclear cells (PBMCs), say, or cell-free plasma (for liquid biopsy).

Cells should be completely lysed to maximize yield (especially important for precious samples), to avoid carryover of contaminants that could inhibit downstream enzymatic steps, and to avoid sequencing biases in relation to other samples.

“There can be significant variation in the extent of degradation depending on storage conditions and age of an FFPE sample,” points out Arvind Kothandaraman, General Manager, Multiomics and Specialty Diagnostics at Revvity. He recommends fine tuning the fragmentation time, and increasing the DNA input, to improve data outcome. “This is important because the effective mass of nucleic acid material participating in the reaction is much lower than with high-quality DNA.”

He adds that it’s also important to measure the integrity of samples before library prep, using an instrument like the LabChip^® GX Touch™ Nucleic Acid Analyzer.

Enrichment

Although NGS costs have plummeted, there are still reasons why researchers might want to limit the amount of DNA that goes into the sequencer. Among other things, the bioinformatics challenges of finding needles in a haystack are significantly less if there’s less hay to sort through. This is especially the case when the starting DNA is damaged and sought-after rare needles (events) are few and far between. Thus researchers will often select the portion of a genome (the exome, for example, or even a more targeted fraction) most likely to contain the mutation of interest.

The same can be said for RNA. Given that mRNA makes up only about 5% of the total RNA in a cell, “the first step in most RNA sequencing workflows is to enrich for transcripts,” Kothandaraman explains. A widely used method is to select for the poly(A) tail found on eukaryotic mRNA using a complementary poly(dT) bait. “However, if a sample is degraded, this approach is unlikely to work. When working with degraded or low-quality RNA, ribosomal depletion is the preferred option for removing ribosomal RNA.”

Repair

FFPE samples display variable, and often substantial levels of DNA damage, as well as yielding low levels of sample suitable for NGS. “One approach to these challenges is to repair the DNA damage ahead of library preparation,” says Fiona Stewart, Associate Director of NGS Portfolio Management at New England Biolabs (NEB).

Several vendors offer enzymes that are specifically designed for FFPE, many of which are available as parts of kits. For example, “we have developed a new mix of DNA repair enzymes that is optimized to repair specific types of damage found in FFPE DNA,” Stewart explains. According to the company’s website the NEBNext FFPE DNA Repair v2 Module will repair deamination of cytosine to uracil; nicks and gaps; oxidized bases; and blocked 3’ ends. “This results in increased yields and quality of libraries, as well as a more streamlined library prep workflow.”

Researchers have also found that short or highly fragmented DNA, such as that found in ancient samples, can benefit from library-preparation methods that process each strand independently as single-stranded DNA.

Streamlining

Every time a human handles a sample there is a chance for cross-contamination, introduction of contaminants, and other errors, let alone the variability associated with manual processing. To address this issue, many vendors have specifically designed kits to reduce the number of steps necessary for preparing an NGS library, by combining multiple enzymatic reactions, for example.

Another option for minimizing manual steps is to automate some or all of them. Labs can choose to use an electronic pipette or liquid-handling robot for one step of the process, or to automate every step from start to finish. It’s important to match the throughput and sample type needs of the lab with the automation platform they choose. Vendors have also optimized kits to accommodate automation—eliminating the need for an ultrasonicator to fragment the DNA, for example.

It is important to ensure that “the reagents and equipment labs choose can accomplish the types of extractions they need to use in their NGS workflows,” Bramlett remarks.