1: Molecular-based methods offer valuable opportunities to assess soil biodiversity in diverse ecosystems. However, their reliability and large-scale applicability depend on developing protocols and establishing high-quality, curated local reference databases. 

2: We focus on soil macroinvertebrates and the assessment of metabarcoding steps, namely the sample decontamination process and the efficiency of taxa recovery, by comparing taxonomy, barcoding, and metabarcoding data. Twenty-four sites were sampled, making up 216 pitfall traps. Following established procedures, specimens were morphologically classified to the lowest taxonomic level. Individuals were then pooled and went through the metabarcoding process. We have developed three experiments: 1) Impact of sample decontamination, where half of the samples were washed with sodium hypochlorite, a commonly used decontaminant. 2) Carabid specimens were identified at the species level by experts, allowing the building of a carabid mock community and a curated reference database for this taxon. The species recovery efficiency was achieved by comparing individual barcoding with metabarcoding samples spiked with DNA from the mock community.

3: Metabarcoding provided a more comprehensive identification of taxonomic orders compared to morphological methods, achieving a higher level of taxonomic resolution. The decontamination process did not show a significant impact on OTU richness, indicating that it may not be essential. Discrepancies in the recovery of carabid species were observed in the three methods.

4: DNA metabarcoding is a promising technique for macroinvertebrate assessment regarding time, efficiency, and costs, while reaching greater depth in taxonomic resolution. The common decontamination step is not crucial for the accuracy of soil macrofauna metabarcoding, and its removal reduces the time and effort required. Here, we demonstrate the potential use of integrative methodologies for robust and rapid biodiversity assessments. Additionally, we state that the reliability of molecular methods is highly dependent on the high quality and availability of well-curated barcode reference databases.