Library Prep for Cell-Free DNA: Preserving Every Fragment

As a first step in many research or clinical tests, obtaining DNA samples from liquid biopsies requires attention to detail. The presence and nature of cell-free DNA (cfDNA) in these samples carries information that is increasingly useful in disease detection and monitoring. Even though cfDNA is present in many kinds of biofluid samples (e.g., blood plasma and serum, urine, saliva, cerebrospinal fluid), preparing cfDNA fragments for sequencing can be challenging due to their smaller size and other factors. The unique qualities of liquid biopsies and cfDNA mean that preparing them for sequencing necessitates particular changes to the conventional library prep methods used for genomic DNA. This article looks at how library prep for cfDNA differs, including efforts to reduce the loss of small fragments through adapter choice, amplification, and careful optimization.

Skip the fragmentation step

Library prep for cfDNA often skips the typical first step. “One fundamental difference is that conventional genomic DNA library preparation often starts by breaking long DNA molecules into smaller fragments, whereas cfDNA preparation is largely about preserving and recovering fragments that are already short,” says Armaan Mehra, Product Manager at Novogene Europe. “cfDNA is often available in limited quantities, so additional fragmentation is therefore generally unnecessary.” This also means size selection must be deliberate, because it can significantly alter the native distribution of fragments (see below). “Overly stringent cleanup can remove shorter fragments, potentially losing valuable tumor-derived signals, while inadequate removal of adapter dimers can compromise sequencing efficiency,” says Mehra.

However, depending upon research interests, researchers may need to enrich for or deplete larger nucleosome-associated fragments. For example, Revvity’s “NEXTFLEX^® Cell Free DNA-Seq Kit 2.0 features an optional magnetic bead-based nucleosome enrichment step, allowing researchers to enrich mononucleosomal fragments or a broader population containing mono-, di-, and trinucleosomal DNA,” explains Pedro Echave, Senior Manager, Global Business Segment at Revvity. “The optional nucleosome enrichment step can also help reduce the representation of high-molecular-weight genomic DNA contamination while retaining the desired cfDNA fragment population.”

Adapters and amplification

Another notable difference is that cfDNAs exist in a variety of fragment lengths, with their ends in many different configurations. “Some are unusually short, nicked, or partially single-stranded, and conventional double-stranded library preparation may underrepresent these populations,” notes Mehra. “Specialized library preparation strategies can improve their recovery, depending on the research application.”

There are many library prep kits already tuned for the unique qualities and low input amounts typical of cfDNA. The shorter length of cfDNA fragments necessitates the use of special adapters, such as stem-loop or bubble adapters, which reduce self-ligation and dimerization. “Our NEBNext^® Ultra™ II DNA Library Prep Kit is highly efficient at converting cfDNA molecules into fully adapter-ligated molecules for sequencing on an array of different platforms,” says Maggie Heider, NGS Development Scientist at New England Biolabs. “We optimize conditions for adapter ligation to improve efficiency and remove unwanted molecules, including excess adapters, before loading the sequencer.”

Because input amounts are often limited, retaining as many fragments as possible using adapters and other optimizations is important throughout the workflow. “We also make use of ultra high-fidelity PCR for amplification of these molecules to ensure the most accurate detection of mutations in patient samples while also maintaining the original fragmentation signatures of these samples,” adds Heider. “As with any library prep, the higher the input amount of material used, the better for maximal unique information gained from the sample, but the library prep can also accommodate trace input levels because sometimes that is all that is available.”

When looking for low-abundance targets, such as circulating tumor DNA or rare variants, it’s essential to track your targets reliably. Many kits incorporate unique molecular identifiers (UMIs), barcodes that tag individual cfDNA fragments prior to PCR amplification, allowing tracking and subsequent error correction if needed. Revvity’s kit “is optimized and validated using 10 ng of cfDNA, making it suitable for applications where starting material is limited, including liquid biopsy research and circulating tumor DNA analysis,” says Echave. “The kit also supports NEXTFLEX Unique Dual Index (UDI) adapters and optional UDI-UMI adapters, providing molecular tracking capabilities for applications requiring rare-variant detection.”

Fragmentomics in cancer

When it comes to DNA, size matters. Most cfDNA fragments are less than 250 bp; in plasma, the amount that might be attached to histones or transcription factors, known as a mononucleosomal fragment, is about 167 bp (147 bp wrapped around the histone octamer core, plus 20 bp attached to a linker histone). In CSF, urine, and saliva biopsies, cfDNA tends to be even more fragmented than in plasma.

But cfDNA exists in a range of sizes, and deviation from the usual range may reflect changes in physiology. In fact, cfDNA is not fragmented randomly. “cfDNA fragmentation patterns carry information about the biological processes and tissues from which it originates,” explains Mehra. “Researchers are increasingly exploring fragmentomics, including fragment-size distributions, DNA end patterns, and nucleosome-associated signatures. These features can provide complementary clues about tissue of origin, chromatin organization, and disease-associated changes, even when detectable mutations are scarce.”

Fragmentation patterns contain information that may help us detect cancer at earlier (and possibly more treatable) stages. “cfDNA is highly fragmented but in a predictable way indicative of release into the bloodstream by apoptotic cells,” notes Heider. For example, plasma from healthy individuals typically shows a peak at 167 bp for mononucleosomal cfDNA fragments. “There is a growing body of data to support additional—in most cases smaller— fragment profiles observed in various disease states, which carry a wealth of information in not only the sizes but the sequence motifs, methylation status, and mutations present in these fragments,” she adds.