Crushing Beats Soaking: Homogenized Insect Samples Deliver Richer DNA Biodiversity Data

DNA metabarcoding has transformed the way ecologists take the pulse of insect communities. Instead of sorting through thousands of specimens under a microscope, researchers can now sequence the genetic traces left in a jar of “biodiversity soup” and generate a species list within days. But behind every metabarcoding result lies a decision that has quietly divided the field: should the entire insect sample be ground up to release its DNA, or can the genetic material be coaxed out gently, leaving the specimens intact for later study? A new study published in Metabarcoding and Metagenomics provides one of the most direct answers yet, and its conclusion favors the blender.

The research team, working with Malaise trap samples collected near the River Boye in Bottrop, Germany, set out to compare the two dominant laboratory approaches side by side. Homogenization, the established standard, physically pulverizes the entire bulk sample before DNA extraction, yielding reliable results but destroying every specimen in the process. Non-destructive lysis, the gentler alternative, incubates whole insects in a buffer containing Proteinase K, an enzyme that digests proteins and liberates DNA through the body wall without mechanical destruction. The appeal is obvious: vouchers remain available for quality control, taxonomic verification, or species description. The question was whether that preservation comes at the cost of missing species.

Earlier validation studies had produced encouraging results for non-destructive lysis, but many relied on artificially assembled mock communities containing far fewer species than a real trap catch. A single Malaise trap sample can hold more than ten thousand insects collected in one week, and mock communities simply cannot reproduce that complexity. Moreover, previous comparisons often treated the two methods with different lysate volumes and extraction chemistries, making it impossible to isolate the effect of the lysis approach itself. The new study addressed this by treating each of five weekly bulk samples with both methods in sequence: first an 18-hour incubation in TNES buffer with Proteinase K at 56 degrees Celsius, then homogenization of the very same specimens, with both fractions processed through identical, highly automated laboratory workflows.

The laboratory pipeline was deliberately rigorous. Twenty-five replicate subsamples were taken from each lysis buffer and from each homogenate, and all liquid handling steps were performed on a Hamilton robot to minimize human variability. DNA was extracted with silica spin columns, amplified in a two-step PCR targeting a standardized barcode region with the fwh primers, and sequenced on an Illumina NovaSeq 6000. Bioinformatic processing through the APSCALE pipeline clustered sequences into operational taxonomic units at a 97 percent similarity threshold, and only reads detected in both technical replicates of each subsample were accepted as genuine signals, guarding against sequencing errors masquerading as rare species.

The results were clear, if modest in magnitude. On average, homogenized samples yielded 3.8 percent more OTUs than their non-destructively lysed counterparts, and in four of the five samples the difference was statistically significant. OTU accumulation curves told a similar story: in three samples, homogenate detected more species even at low replicate numbers, and adding more replicates never brought the non-destructive lysis curves up to the homogenate level. Intriguingly, the OTUs missed by non-destructive lysis did not belong to any particular heavily sclerotized lineage; instead, exclusive detections in both treatments tended to come from families with small body sizes, suggesting that low biomass, rather than tough exoskeletons, governs whether a specimen releases enough DNA to be detected.

The mechanistic explanation follows directly from the chemistry. Homogenization places far more cells in contact with the lysis buffer, so more DNA molecules enter the extraction. Non-destructive lysis, by contrast, depends on DNA diffusing out of intact bodies, and specimens with little tissue to begin with may simply never release a detectable quantity. A dilution effect likely compounds the problem: when the concentration of DNA in the lysate is low, OTUs hovering near the detection limit of the metabarcoding workflow fall below the threshold of reliable recovery. The authors also note that some DNA leaches into the buffer during the 18-hour soak and is lost before homogenization, meaning that homogenization performed alone might recover an even richer community than the sequential design of this experiment could demonstrate.

There was, however, a reassuring counterpoint. When the researchers examined community composition across the five consecutive weekly samples using Jaccard beta diversity, both methods painted the same ecological picture. The ordination plots showed that samples separated primarily by week of collection, not by treatment, meaning that either method would detect the same seasonal turnover patterns. Between 80 and 84 percent of OTUs were shared between the two treatments in every sample, and the authors conclude that the two approaches can be intercalibrated and used in standardized biodiversity assessments, provided the goals of a study are matched to the strengths of each method.

Yet the study also delivered a sobering cost-benefit verdict. Non-destructive lysis is expensive: the quantities of Proteinase K required for a single bulk sample run approximately 40 to 60 euros depending on sample biomass, which alone equals the cost of the entire routine analysis in the authors’ established workflow. The method also demands more manual labor and longer handling times, and because extracting the full lysate volume would be needed to maximize species recovery, automation becomes complicated or impossible. After the 18-hour incubation, the specimens were already so softened that individual handling for morphological identification was extremely limited, undermining the very rationale for preserving them. Shorter incubations would protect morphology better but would sacrifice DNA yield, creating an unavoidable trade-off.

The authors therefore conclude that non-destructive lysis is not suitable for integration into automated, standardized monitoring protocols in the foreseeable future. They acknowledge that earlier studies reached more optimistic conclusions but argue that methodological inconsistencies, such as unequal sample volumes and different extraction methods between treatments, limited the comparability of those comparisons. There is also a deeper conceptual problem: even when non-destructive lysis succeeds in preserving specimens, metabarcoding cannot link the detected OTUs to individual bodies, so researchers hoping to confirm a species of interest would still face the daunting task of sorting through thousands of insects to find it.

Where does this leave the field? For large-scale biomonitoring programs that process thousands of Malaise trap samples, homogenization remains the more comprehensive, cheaper, and automation-friendly choice, and the modest 3.8 percent richness gain compounds into substantial additional data at scale. Non-destructive lysis retains genuine value in niche applications where morphological integrity is paramount and budgets are generous, such as museomics projects analyzing irreplaceable historical specimens. The study’s most practical contribution may be its demonstration that the two methods produce comparable community patterns despite differing in absolute richness, giving monitoring programs a defensible basis for combining data across protocols while choosing the destructive route whenever the specimens themselves are expendable.

Subject of Research: Comparison of homogenization and non-destructive lysis for DNA metabarcoding of insect bulk samples

Article Title: Homogenization of insect bulk samples yields more comprehensive yet comparable biodiversity data than non-destructive lysis

Article References: Wolany, L., Enss, J., Leese, F., & Buchner, D. (2024). Homogenization of insect bulk samples yields more comprehensive yet comparable biodiversity data than non-destructive lysis. Metabarcoding and Metagenomics, 8, Article e129814. https://doi.org/10.3897/mbmg.8.129814

Image Credits: AI Generated

DOI: 10.3897/mbmg.8.129814

Keywords: DNA metabarcoding, insect biodiversity, Malaise traps, non-destructive lysis, homogenization, biomonitoring, operational taxonomic units, Proteinase K, bulk samples, beta diversity, molecular ecology, insect