Rapid Identification of Edible Insect Species in Food Using MALDI‐TOF Mass Spectrometry

ABSTRACT

Rationale

Edible insects are emerging as sustainable, nutritious ‘foods of the future’ and are gradually introduced to the European market as novel foods. Ensuring consumer safety and preventing fraud requires legal regulation, which in turn depends on reliable analytical methods. Sensitive, rapid techniques capable of identifying insect species are needed to support enforcement and monitoring of legislation across different food matrices.

Methods

Powders of 10 insect species, including all four edible insects authorised as novel foods in the European Union, were digested with trypsin without toxic extraction agents. Samples were purified using ZipTip C₁₈ and analysed by matrix‐assisted laser desorption/ionisation time of flight mass spectrometry (MALDI‐TOF MS) in positive reflector mode. Species‐specific m/z values were identified using a simple structured query language‐based search. Six model mixtures and 10 commercial products were tested with the same workflow to assess genus‐ and species‐level authentication.

Results

Six model mixtures and 10 commercial products were authenticated using previously created genus‐ and species‐specific m/z databases. Model mixtures with dominant insect proteins allowed correct genus‐ and species‐level identification (7–20% peptide matches). Closely related species or dominant spectral components increased the misidentification. In the commercial products, identification was adequate for high‐insect‐content samples, but species assignment was ambiguous or incorrect in products with low insect content (7–10%) and high plant protein content.

Conclusions

MALDI‐TOF MS enables rapid genus‐level insect identification and species‐level discrimination, especially when dealing with unprocessed single‐species powders. Accurate species assignment depends on insect abundance and matrix complexity of the inspected products. As well, the use of MALDI‐TOF MS for species identification of products with low insect content or high plant protein is still limited and may require complementary methods such as LC–MS/MS to achieve unambiguous species identification.

1. Introduction

Foods made from edible insects offer sustainable and cost‐effective production, high nutritional value and functional properties that improve food quality. When added to traditional foods in appropriately selected amounts, they are also positively evaluated in terms of sensory properties [1, 2].

Consuming edible insects has a thousand‐year tradition in many cultures. Currently, over 2300 insect species are consumed worldwide [3]. However, this practice has not historically been common in the European Union. European legislation regards edible insects and insect‐derived products as so‐called novel food in accordance with Regulation (EU) 2015/2283 [4]. Each such novel food, including foods derived from edible insects, must undergo an authorisation process before being placed on the European market. The initiation of the authorisation process may be undertaken either on the Commission's initiative or following an application to the Commission by an applicant (e.g., a food business operator). The decision on the particular novel food approval is then granted by the Commission based on the scientific opinion issued by the European Food Safety Authority [4]. To date, four edible insect species have been authorised for human consumption on the European market: adult Acheta domesticus based on Regulations (EU) 2022/188 [5] and (EU) 2023/5 [6], adult Locusta migratoria based on Regulation (EU) 2021/1975 [7], Alphitobius diaperinus larvae based on Regulation (EU) 2023/58 [8] and Tenebrio molitor larvae based on Regulations (EU) 2021/882 [9], (EU) 2022/169 [10] and (EU) 2025/89 [11]. The corresponding Commission Implementing Regulations [5, 6, 7, 8, 9, 10, 11] specify much more detailed requirements for the form of the products, inorganic and microbiological contaminants, the insects' fasting period before killing, etc. The handling of edible insects is also subject to other European regulations, such as (EU) 1069/2009 [12], which establishes that edible insects fall under the category of farmed animals and are therefore subject to strict feed requirements, (EU) 852/2004 [13] on hygiene requirements and (EU) 1169/2011 [14] on proper food labelling. The import of edible insects into the EU countries is also regulated, by Regulation (EU) 2021/405 [15]. Compliance with the national legislation of individual EU member countries must also be ensured.

Some recently published scientific studies [16, 17] report real cases of mislabelling of insect‐based foods, including adulteration through substitution with undeclared or even unauthorised insect species and also point to the potential adulteration of insect‐based foods through partial substitution with much cheaper plant‐derived proteins from soybean [16, 17]. This highlights the importance of developing new analytical methods and protocols for reliable monitoring of compliance with the above‐mentioned legislation, thereby protecting consumers from fraud and health risks.

The development of new methods for the control of insect‐based foods is being investigated by experts in microscopy [18], spectroscopy [19, 20], histochemistry [21], immunochemistry [22, 23], genetics [24, 25, 26, 27, 28, 29], lipidomics [30, 31, 32], metabolomics [32, 33, 34, 35, 36] or proteomics [32, 37, 38, 39, 40, 41]. Especially for the last two mentioned fields, mass spectrometry is a key tool.

Metabolome analysis can be performed using techniques such as liquid chromatography–tandem mass spectrometry (LC–MS/MS) [33] or direct analysis in real‐time high‐resolution mass spectrometry (DART‐HRMS) [34, 35, 36]. Using DART‐HRMS, distinct lipid and purine fingerprints reflecting the age group and diet of T. molitor were identified. Comparison of metabolomic and metagenomic data revealed low levels of certain fatty acids associated with a reduced abundance of Enterococcus in Tenebriones fed with chicken feed [36]. Tata et al. [35] used DART‐HRMS coupled with mid‐level data and a learning method for distinguishing among A. domesticus, Bombyx mori, Hermetia illucens and T. molitor according to their metabolic fingerprints. The results showed the high potential of this technique to identify adulteration via complete substitution [35]. However, Poma et al. [34] call for caution when using multivariate statistical approaches, as well as warn against the challenges of metabolomics such as challenging sample preparation (especially in case of targeting compounds present in lower concentrations), difficulty of data evaluation and interpretation due to their complexity. Analyte instability over time is another obstacle [42, 43].

For species authentication purposes advantageously, proteomics targets analytes that are stable [42, 43]—mainly structural proteins and peptides [44, 45]. Moreover, the high abundance of proteins in insects (approximately 35–60%, based on dry matter) [46] makes the proteomic approach a potential method of choice for edible insect‐based food authentication, serving as a possible alternative to more conventional DNA analysis [24, 25, 26, 27, 28, 29]. Francis et al. [37] also note that DNA approaches could preferentially target DNA sequences of proteins previously identified by MS techniques to develop specific real‐time PCR tests [37]. LC–MS/MS techniques are considered promising for species identification and even for detecting allergens and putative allergens [39]. Francis et al. [37] show the species‐specificity of edible insect proteomes but also mention the limited sequence coverage in public reference databases [37]. To overcome this obstacle, researchers try to develop methods less dependent on database quality, such as direct MS/MS spectrum comparison with spectral library generation [41] or homology‐based [39] proteomics. Another widely used proteomic technique that is largely independent of sequence data in reference databases and which has been used in the past for many discriminative studies [47, 48, 49, 50, 51, 52] of insects, matrix‐assisted laser desorption/ionisation—time of flight mass spectrometry (MALDI‐TOF MS), has also been used for distinguishing edible insect species [40]. In the cited study [40], four edible insect species approved for the European market ( A. domesticus, A. diaperinus, L. migratoria and T. molitor ) were analysed for mutual interspecific discrimination.

Also in the presented study, MALDI‐TOF MS was used for distinguishing of four edible insect species approved in EU for human consumption. However, an entirely different strategy targeting peptides instead of proteins was adopted. Moreover, the range of analysed insect species was expanded for the purpose of providing better context for the obtained data. In total, 10 insect species ( A. domesticus, A. diaperinus, Blaptica dubia , G. assimilis, H. illucens, L. migratoria, Schistocerca gregaria, Shelfordella tartara, T. molitor and Zoophobas morio ) were selected to identify their discriminatory peptide markers. The selection included all insect species authorised in the EU for human consumption, several species authorised for use in feed, as well as additional insect species with potential for future authorisation. Powders from all mentioned insect species were subjected to direct enzymatic digestion with trypsin, known for its high specificity. No previous protein extraction has been used [53, 54, 55]. The presented protocol thereby avoids the excessive use of toxic reagents, which is in accordance with the notice in Tata et al. [35]. The evaluation strategy was also different from Ulrich et al. [40]. Instead of a software‐based comparison of entire spectra, so‐called peak lists (manually selected m/z values) were compared using the PostgreSQL database system, the pgAdmin interface and an internal protocol [56] to find species‐specific m/z values functioning further as markers. The applicability of the found markers was then tested for insect species identification in six model blended insect powders and 10 commercial edible insect‐based foods. Both, protocol from Ulrich et al. [40] and the protocol presented in this paper appear to become useful sustainable approaches. Further optimisation of these protocols and their potential integration with the simultaneous identification of microorganisms in the sample (which is already routinely performed using MALDI‐TOF MS) [57, 58, 59] could strengthen MALDI‐TOF MS as an effective analytical tool providing both microbiological and authentication functions, which are, inter alia, required by the applicable above‐mentioned European regulations.

2. Materials and Methods

2.1. Reagents and Materials

Acetonitrile (LC–MS grade), 2,5‐dihydroxybenzoic acid (DHB) (≥ 99.5% (HPLC), ultra pure) and trifluoroacetic acid (TFA) (98%) were purchased from Sigma‐Aldrich (USA). Ammonium hydrogen carbonate (99.5%) was obtained from Fluka (Honeywell Research Chemicals, Germany). Peptide Calibration Standard II was purchased from Bruker Daltonics (Germany). Pierce Trypsin Protease MS Grade was obtained from ThermoFisher Scientific (USA). The commercially available reverse‐phase C₁₈ ZipTip pipette tips were purchased from Merck Life Sciences (USA). The water was purified with a Milli‐Q water purification system from Merck Life Sciences (USA).

2.2. Analysed Samples

2.2.1. Reference Samples of Insects

Ten species of insects (adults of A. domesticus, B. dubia , G. assimilis , H. illucens , L. migratoria, S. gregaria, S. tartara , and larvae of A. diaperinus, H. illucens , T. molitor and Z. morio ), along with their homogenised powders, were analysed. The samples were either reared in the insectarium at the Czech University of Life Sciences Prague or provided by 12 different farmers from the Czech Republic and delivered by PAPEK s.r.o., Czech Republic. All insect specimens were starved for at least 24 h before killing, killed by freezing at −20°C or by boiling in water for 5 min, freeze‐dried (L10‐55 PRO lyophilizer, Gregor Instruments s.r.o., Sazava, Czech Republic), cryogenically homogenised in the presence of liquid nitrogen using a laboratory knife mill (A11 Basic Analytical mill, IKA Labortechnik, Staufen im Breisgau, Germany) and stored in a freezer (−80°C). These primary samples were later sampled for measurement on the day of analysis. For analysis, two amounts (2.0 mg and 5.0 ± 0.2 mg) of each sample were weighed in triplicate. The genetic data, derived from the same research project (QK23020101, Ministry of Agriculture of the Czech Republic), confirmed the presence of only the tested species in the samples.

2.2.2. Model Mixtures

Six mixtures (M) of different insect powders (reference samples from Section 2.2.1) were mixed in mass ratios and prepared as single‐blind samples. For analysis, 5.0 ± 0.2 mg of each sample was weighed in three repetitions.

M1. Z. morio: T. molitor : A. diaperinus (1:1:1)

M2. T. molitor : S. gregaria (1:1)

M3. G. assimilis : A. domesticus (1:1)

M4. L. migratoria: S. gregaria (1:1)

M5. G. assimilis : L. migratoria (1:1)

M6. B. dubia: A. domesticus (1:1)

2.2.3. Commercial Food Products

Ten commercial products (CPs) were analysed. Their compositions, including the declared insect species and their proportions, are presented in Table 1.

TABLE 1.

Summary of analysed commercial products with their ingredient composition.

Sample	Product name	Producer	Declared insect	% of insect	Labelled composition
CP1	Crispy crickets horseradish	Grig	Acheta domesticus	93	Dried house crickets ( A. domesticus ) 93%, rapeseed oil 2%, spice mix 5% (salt, dried whey, horseradish, sugar, yeast extract, smoke flavour, onion, paprika, aroma E635)
CP2	Cricket powder	Grig	A. domesticus	100	Cricket powder ( A. domesticus ) 100%
CP3	Chili grasshoppers	Grig	Locusta migratoria	86	Dried migratory locusts 86% (L. migratoria ), sunflower oil 4%, chili, salt
CP4	Pepper insect snack (Tenebrio)	Essento	Tenebrio molitor	67	Dried mealworm ( T. molitor) 67%, pink pepper 12%, glucose, sunflower oil, salt 4.4%, black pepper 0.9%
CP5	Worms salted caramel	Grig	T. molitor	97	Dried mealworm larvae ( T. molitor ) 97%, flavouring mix 2% (sugar, edible salt, aroma, spices, colouring: ammonium caramel)
CP6	Parmesan worms	Grig	T. molitor	97	Dried mealworm larvae ( T. molitor ) 97%, flavouring mix 2% (glucose syrup, flavour enhancer: monosodium glutamate, edible salt, aroma, yeast extract, colouring [beet red], spice extracts), edible salt 1%
CP7	Salted worms	Grig	T. molitor	99	Dried mealworm larvae ( T. molitor ) 99%, edible salt 1%
CP8	Cricket triangles tomato and basel	Grig	A. domesticus	7	Gluten‐free chickpea flour, rice flour, cricket powder ( A. domesticus ) 7%, potato starch, rapeseed oil, flavouring ingredient (dextrose, salt, tomato powder, yeast extract, garlic, paprika, basil, spices, spice extracts, aroma), salt
CP9	Pea cricket protein chips, poppy seeds and salt	Sens	A. domesticus	10	Pea flour (76%), cricket flour ( A. domesticus) (10%), poppy seeds (7%), sunflower oil, corn flour, sea salt (2.4%), carrot, onion, yeast extract, parsley, spice mix
CP10	Strawberry cricket protein blend	Sens	A. domesticus	10	Pea protein, cricket powder ( A. domesticus ) 10%, sunflower protein, strawberry powder, natural strawberry flavour 2%, dye—beetroot powder, thickener—guar gum, sunflower lecithin, chicory fibre, sweetener sucralose. Note: insects killed by cooling

2.3. Sample Preparation

At first, the optimal sample treatment conditions were determined. Optimising the sample preparation procedure consisted of finding the proper amount of sample needed for analysis, the duration of trypsin digestion and finding the appropriate sample‐to‐matrix (DHB) ratio and purification of the samples using ZipTip C₁₈.

The weighed‐out samples (2.0 ± 0.2 and 5.0 ± 0.2 mg) were digested in 10 μL, respectively, 25 μL of 50 mM NH₄HCO₃ containing 0.02 mg/mL of trypsin (EC 3.4.21.4), further diluted with another 30 μL, respectively, 50 μL of 50 mM NH₄HCO₃ at 37°C with constant shaking for 2, 4, 6 or 18 h. The enzymatic cleavage was terminated by adding 10% TFA solution to a final concentration of 0.5% TFA. After the tryptic digestion, the samples were purified and concentrated on reverse‐phase ZipTip C₁₈. The samples of model mixtures and commercial food products (5.0 ± 0.2 mg) were prepared for analysis in the same way.

After the purification process, 10 μL of each purified sample was obtained and mixed with the DHB matrix solution (8.5 mg of DHB in 0.5 mL of a mixture of acetonitrile/0.1% TFA in water) in 1:3 and 1:5 ratios.

2.4. MALDI‐TOF MS Measurements and Data Acquisition

Two microlitres of the resulting mixture of the sample and DHB solution was deposited in triplicate onto a stainless steel MALDI target and allowed to air‐dry at ambient temperature. Calibration spots were prepared by mixing 1 μL of the calibration standard (Bruker Daltonics) with 5 μL of DHB matrix. Calibration was carried out before the measurement and subsequently after every fifteenth spot during the analysis. Measurements were performed using an Autoflex Speed MALDI‐TOF mass spectrometer (Bruker Daltonics, Germany). Analyses were carried out in positive reflector mode using a Nd:YAG laser operating at 355 nm with an energy of 50–60%. Spectra were acquired using FlexControl (Bruker Daltonics, Germany) within a mass range of 680–5000 m/z. High voltage settings: Ion Source 1, 19.0 kV; Ion Source 2, 16.5 kV; lens, 8.8 kV; reflector 21.0 kV; Reflector 2, 9.45 kV; pulsed ion extraction set to 110 ns. The signal was suppressed until 450 Da. The corresponding spectrum was manually obtained from a total of 10 000 shots (10 × 1000 shots) per spot.

2.5. Searching for Peptide Markers—m/z Values

For the search of peptide markers only spectra obtained from samples of insect specimens killed by cooking were used. The peaks occurring in the spectra of all samples of one species but in no spectra of samples of other insect species were searched to distinguish among the analysed insect species from one another. Such peptides (described by m/z values) can be considered as species‐ or genus‐specific markers.

Initially, all acquired mass spectra were processed in mMass (Version 5.5.0). After spectrum smoothing and baseline correction, peaks were manually selected. All selected peaks met the criterion of a signal‐to‐noise ratio greater than 3. Approximately, 80–130 peaks meeting the above‐mentioned criterion were selected in each spectrum. The resulting peak lists (lists of the selected m/z values) for each sample were exported to Microsoft Excel and further processed in PostgreSQL via the pgAdmin 4 interface. Data evaluation using PostgreSQL was performed following a specially designed internal protocol [56]. The search for m/z values that repeated in the spectra with a certain frequency (which is adjustable) took place first. This search was carried out for each insect genus (at first) and later for species separately. Its purpose was to obtain values that occur in every spectrum obtained from the samples of a particular genus or species. In our case, the frequency was set to nine; that is, the given m/z value had to be repeated in nine out of nine spectra (three sample replicates, three spots on the MALDI plate from each sample replicate within one genus/species).

The usually used automatic software‐based MSP (Main Spectra Profile) approach [40] was not employed, as the aim was to develop an alternative method free from paid commercial software dependence.

3. Results and Discussion

Measurements using MALDI‐TOF MS were performed on all 10 insect species. The results represent peptide markers (m/z values) specific to each genus or species analysed in this work. The markers were subsequently used for inspection of six model mixtures and 10 CPs. As mentioned above, although the well‐established MSP software‐based approach was not employed in this study, an alternative workflow was used. In the MSP approach, spectra and peak intensities are typically averaged to generate reference profiles. In contrast, the method used in this study, implemented using a PostgreSQL database created in‐house, does not rely on spectral averaging. Instead, only peptides detected in a defined number of spectra (e.g., 9 out of 9) are included in subsequent analyses; these parameters can be adjusted if necessary (e.g., 8 out of 9 spectra).

In addition, peak intensities are not averaged, which helps to minimise the influence of peaks present only in a small subset of spectra. In this study, strict criteria were applied for marker selection, requiring the consistent presence of the given m/z value in all spectra of the respective insect species or genus and its complete absence in spectra of all other analysed species. This approach supports robust marker identification and reduces the likelihood of artefacts (such as contamination).

3.1. Sample Treatment Optimisation

To obtain high‐quality data, the search for suitable sample pretreatment conditions was crucial. The tested conditions were 2 and 5 mg (±0.2 mg) sample weights, a duration of tryptic cleavage (2, 4, 6 or 18 h), the sample‐to‐matrix ratio (1:3 and 1:5) and the method of insect killing (boiling or freezing). For the test, two insect genera were chosen: T. molitor (killed by boiling) and G. assimilis (killed by freezing).

The spectra obtained from the insect killed by freezing, compared with the spectra of the insect killed by boiling, provided a lower number and intensities of peaks and showed a lower ratio of the signal‐to‐noise (Figure 1). The killing method significantly affected protein integrity and the resulting spectral quality. The presented results provide further evidence supporting the validity of previously published claims by Leni et al. [60]. As they mention, amino acids Cys and Lys decrease during killing by freezing due to their involvement in the process of melanisation which occurs enzymatically during freezing. Thus, protease trypsin, which specifically cleaves immediately C‐terminal to arginine and lysine residues, unless followed by proline, loses available cleavage sites in the substrate. On the other hand, boiling causes protein denaturation, partial inactivation of endogenous enzymes and release of proteins from protein‐chitin complexes leading to more reproducible and higher‐quality MALDI‐TOF MS spectra [60].

FIGURE 1.

Summed mass spectra of Tenebrio molitor killed by boiling or freezing reveal variations in number of detected peaks and their intensities.

The highest number of peaks was found in spectra of samples cleaved for 18 h, followed by those digested for 6 h. Although a longer digestion time resulted in slightly higher information content, a digestion time of 6 h was selected to accelerate the analysis and thus reduce the cost of analysis. Shorter sample preparation and, consequently, lower operational costs were prioritised, as the reduction in spectral information was negligible. During the experiment, high‐quality spectra could not be obtained for A. domesticus samples (2 mg) mixed with DHB at a 1:3 ratio after 6 h of digestion; nevertheless, a digestion time of 6 h was retained based on the satisfactory results achieved for T. molitor samples (5 mg) prepared at a sample‐to‐matrix ratio of 1:5. Based on the results summarised in Table S1, the optimal conditions for sample preparation and subsequent peptide marker identification were determined as follows: insects killed by boiling, a sample weight of 5 mg, 6 h of enzymatic digestion and a sample‐to‐DHB matrix ratio of 1:5.

3.2. Searching for Genus/Species‐Specific Markers

The genus‐ or species‐specific values are m/z values that occurred in one genus (or species), but not in any of the others analysed. With regard to the later evaluation of the samples, firstly, the reference samples of insects killed by boiling were evaluated in order to obtain characteristic values for distinguishing individual genera. Thus, m/z values from the following genera groups were compared: black soldier fly ( H. illucens ), crickets ( A. domesticus , G. assimilis ), cockroaches ( B. dubia , S. tartara ), locusts ( L. migratoria , S. gregaria ) and darkling beetles—the family of Tenebrionidae ( A. diaperinus , T. molitor , Z. morio ). The complete list of unique values for all five groups of genera is shown in Table S2. The most characteristic values enabling genus‐level discrimination were found for the group of cockroaches (264); approximately 2.5 times fewer were found for locusts (107) and darkling beetles (105), 3.3 times fewer for crickets (79), and the lowest number was observed for the black soldier fly (50).

In the next step, after the discrimination of insect genus, the insect species was looked for in each genus. The list of these characteristic m/z values is shown in Tables S3–S6. To discriminate between the cricket species A. domesticus and G. assimilis , 143 characteristic values and 130 were found, respectively. For mutual discrimination of cockroaches, 147 values for B. dubia and almost twice as many, 289, for S. tartara , and 171 and 147 values for locusts, L. migratoria and S. gregaria , were revealed. For darkling beetles, specifically, 84 values for A. diaperinus , 130 values for T. molitor and 113 values for Z. morio were found.

3.3. The Identification of Insect Genus and Species in Model Mixtures

Six model mixtures in different ratios of insect powder of selected genera or species were analysed in three repetitions (A, B and C). The m/z values selected from mass spectra were compared with the database containing genus‐specific m/z values (Table S2) and consequently to the database with species‐specific m/z values (Tables S3–S6). The ratio of all detected m/z values to species‐specific markers for genus/groups of crickets ( A. domesticus and A. assimilis ), darkling beetles ( Z. morio, T. molitor and A. diaperinus ) and locusts ( L. migratoria and S. gregaria ) ranged approximately from about 7% to 20% (Table 2). The percentage indicates the ratio between characteristic peptides for genus/insect groups and all peptides detected in the sample spectrum. Although the percentage of matches may appear low, it is consistent with previously reported findings, where matches for determining the animal origin of blood in artworks ranged from 10% to 28% [61]. The identification of all insect species at the level of genus and species was correct in four samples: M1 (20.4 ± 0.8% for the group of darkling beetles), M2 (12.1 ± 1.0% for darkling beetles and 7.5 ± 2.3% for locusts), M3 (8.9 ± 1.8% for crickets) and M4 (17.0 ± 1.2% for locusts). In samples M5 (9.3 ± 1.8% for crickets and 6.9 ± 0.9% for locusts) and M6 (15.4 ± 0.9% for cockroaches and 6.8 ± 1.8% crickets), the identification of both genera of insects was successful, but the determination of the species of one of the insects was incorrect. In sample M5, S. gregaria was identified instead of L. migratoria . The ratio of characteristic peptides for S. gregaria was equal (Repetition A) to L. migratoria or even higher (Repetitions B and C). In the case of mixture M6, the incorrect determination of cricket species could be caused by the presence of cockroach, because cockroaches provide many peaks in mass spectra and some of them could mask as specific m/z values in the cricket species distinguishing. The distinguishing of species is done using the whole set of m/z values found in the mass spectra.

TABLE 2.

Numbers of all m/z values found in mass spectra of model mixtures made of insect powders/number of found specific m/z values for the given genus/species (in bold). The ratios in table in section about species determination are: cockroaches: Blaptica dubia : Shelfordella tartara , crickets: Gryllus assimilis : Acheta domesticus , darkling beetles: Zophobas morio : Tenebrio molitor : Alphitobius diaperinus , and locusts: Locusta migratoria : Schistocerca gregaria .

Model mixture	Repetition					Repetition			Identified species	Correct identification
	A	B	C	Average % of matches ± SD	Identified genus	A	B	C
M1	134/26	157/32	122/26	20.4 ± 0.8	Darkling beetles	14:20:7	13:27:8	9:22:9	Z. morio : T. molitor : A. diaperinus	Yes
M2	123/15, 6	111/12, 8	152/20, 16	12.1 ± 1.0/7.5 ± 2.3	Darkling beetles and locust	Darkling beetles: 3:26:0 Locust: 15:26	Darkling beetles: 1:22:2 Locust: 10:22	Darkling beetles: 4:28:1 Locust: 18:29	T. molitor : S. gregaria	Yes
M3	245/27	149/10	99/9	8.9 ± 1.8	Crickets	Crickets: 34:33	Crickets: 18:14	Crickets: 25:13	G. assimilis : A. domesticus	Yes
M4	138/25	137/24	183/28	17.0 ± 1.2	Locusts	19:28	21:30	23:42	L. migratoria : S. gregaria	Yes
M5	156/11, 10	97/9, 8	147/17, 9	9.3 ± 1.8/6.9 ± 0.9	Crickets and locusts	Crickets: 31:16 Locust: 15:15	Crickets: 20:15 Locust: 13:16	Crickets: 27:13 Locust: 10:22	G. assimilis : S. gregaria	Both genera correct; species of locust is incorrect
M6	209/32, 9	195/28, 15	168/28, 14	15.4 ± 0.9/6.8 ± 1.8	Cockroaches and crickets	Cockr.: 43:7 Crickets: 23:1	Cockr.: 39:4 Crickets: 21:14	Cockr.: 40:8 Crickets: 22:15	B. dubia : G. assimilis	Both genera correct; species of cricket is incorrect

3.4. The Identification of Insect Genus and Species in Commercial Food Products

Ten commercial food products (CPs) were analysed with the aim of identifying the insect at the level of genus and species (Table 3). Seven products (CP1–CP7) contained only insects and spices, which could contain small amounts of proteins from dried whey or yeast extracts, and other non‐proteinaceous components. Samples CP8, CP9 and CP10 contain mainly proteinaceous components like pea protein, chickpea and rice flour. The insect ( A. domesticus ) content was only 7% in CP8, respectively, 10% in CP9 and CP10. The exact composition declared by manufacturer is reported in Table 1.

TABLE 3.

Insect identification in commercial food products. The ratios in the table in the section about species determination are: crickets: Gryllus assimilis : Acheta domesticus , darkling beetles: Zophobas morio : Tenebrio molitor : Alphitobius diaperinus , and locusts: Locusta migratoria : Schistocerca gregaria .

Sample	No. of peaks/no. of matches with genus	% of matches with genus	Identified genus	Ratios of no. of matches for species	Identified species	Correct identification
CP1	17/3	17.7	Crickets	0:5	A. domesticus	Yes
CP2	91/9	9.9	Crickets	0:26	A. domesticus	Yes
CP3	78/7	9.0	Locusts	16:11	L. migratoria	Yes
CP4	61/5	8.2	Darkling beetles	0:7:1	T. molitor	Yes
CP5	62/14	22.6	Darkling beetles	0:19:0	T. molitor	Yes
CP6	89/6	5.3	Darkling beetles	0:10:0	T. molitor	Yes
CP7	120/20	16.7	Darkling beetles	0:31:0	T. molitor	Yes
CP8	117/4, 6, 6	3.4; 5.1; 5.1	Crickets, cockroaches, locusts	—	—	No
CP9	128/5	3.9	Crickets	6:3	G. assimilis	Genus—yes Species—no
CP10	124/7, 10, 7, 4	5.6; 8.1; 5.6; 3.2	Crickets, cockroaches, darkling beetles, locusts	—	—	No

According to Table 3, the insect species were successfully identified in the first seven products (CP1–CP7), which contain insects as the major source of proteins. At the genus level, the highest match rates selected the best value from the three replicates were 17.7% for crickets in CP1 (Crispy crickets horseradish, 93% of A. domesticus ), 22.6% for darkling beetles in CP5 (worms salted caramel, 97% of T. molitor ) and 9% for locusts in CP3 (chili grasshoppers, 86% of L. migratoria ). Clear species‐level identification was achieved for all samples CP1–CP7 containing whole insect bodies or prepared as powder (CP2, cricket powder, 100% of A. domesticus ). Proteomic similarity between locust species was observed in sample CP3, where a higher number of species‐specific peptides was detected for L. migratoria than for S. gregaria (16:11). Identification of T. molitor within the darkling beetle group, which includes three insect species— Z. morio , T. molitor and A. diaperinus—was unambiguous in all four samples CP4–CP7, regardless of the insect content, which ranged from 67% to 99%. Based on these results, flavouring ingredients do not have a significant impact on the correct determination of insect species.

The last three samples (CP8, CP9 and CP10) contained many plant‐derived ingredients; for example, sample CP9 declared 76% of pea flour. In sample CP9 (Pea cricket protein chips containing 10% A. domesticus ), the cricket genus was correctly identified; however, G. assimilis was detected instead of A. domesticus . The ratio of peptides indicating the presence of G. assimilis and A. domesticus was 6:3. The incorrect identification of the insect species may have been influenced by the high proportion of the above‐mentioned pea flour.

In samples CP8 (cricket triangles tomato and basel, 7% A. domesticus ) and CP10 (strawberry cricket protein blend, 10% A. domesticus ), insect species could not be unambiguously identified due to the predominance of vegetal proteins (chickpea and rice flour, sunflower protein, pea protein). In CP8, crickets, cockroaches and locusts were detected, with matches of 3.4%, 5.1% and 5.1%, respectively, whereas in CP10, crickets, cockroaches, darkling beetles and locusts showed match percentages of 5.6%, 8.1%, 5.6% and 3.2%, respectively. The black soldier fly ( H. illucens ) was correctly not detected in both products.

The limited species‐level identification likely resulted from suppression of less abundant insect‐derived peptides (the declared content of insect in food products was from 7% to 10%) or from overlapping m/z values between vegetal and yeast extract peptides and insect peptides. The m/z range used for spectral comparison was 680–5000, with the most detectable peptides falling between 900 and 2000 m/z, allowing abundant vegetal/yeast extract peptides to share m/z values with insect peptides. These limitations can be overcome by LC–MS/MS [45, 61] which provides additional peptide sequence information alongside m/z values, enabling more reliable species identification. Furthermore, identification in CP10 may have been affected by the lethal freezing of the crickets, as non‐boiled specimens generally yield lower‐quality mass spectra (Figure 1) due to above‐mentioned reasons supported/described by Leni et al. [60].

3.5. MALDI‐TOF MS Performance in Insect Identification Across Different Food Matrices

The analysis of model mixtures (M1–M6) and CPs (CP1–CP10) demonstrated that MALDI‐TOF MS enables reliable insect identification at the genus level and, under favourable conditions, also at the species level. In model mixtures M1–M4, correct genus‐ and species‐level identification was achieved for all insects present despite low proportions of genus‐ or species‐specific peptides (approximately 7–20%, Table 2), indicating that a limited number of characteristic peptides is sufficient for unambiguous assignment when insects represent the dominant protein component. Similarly, in CPs CP1–CP7, where insects constituted the main source of proteins, MALDI‐TOF MS enabled correct genus‐ and species‐level identification regardless of insect form and the presence of flavouring ingredients (Table 3). In particular, T. molitor was consistently distinguished from other darkling beetles in samples CP4–CP7 across a wide range of insect contents (67–99%).

The limitations of the method were apparent in samples containing closely related species or low proportions of insect proteins. In model mixtures M5 and M6, both insect genera were correctly detected; however, species‐level identification of one component was incorrect, likely due to comparable ratios of species‐specific peptides ( L. migratoria vs. S. gregaria in M5) or spectral interference from a dominant component (cockroach proteins in M6). In CPs CP8–CP10, where insects accounted for only 7–10% and plant‐derived proteins predominated, species‐level identification was ambiguous or not possible and multiple insect genera were detected at low match percentages. Overall, these results indicate that MALDI‐TOF MS is well suited for rapid screening and reliable genus‐level identification, whereas species‐level discrimination is dependent on insect abundance and matrix complexity and may require complementary analytical techniques. Despite the high cost of instrumentation, the demonstrated performance of MALDI‐TOF MS supports its use in centralised or shared analytical facilities as part of multi‐method strategies for insect authentication.

4. Conclusions

In this paper, first, the suitable sample preparation conditions for edible insects were determined: to analyse insects killed by boiling instead of freezing, a sample weight of 5 mg, 6 h of enzymatic digestion and a sample‐to‐matrix ratio of 1:5.

Using above mentioned sample pretreatment conditions and subsequent processing of the obtained MS data in the PostgreSQL database system using the pgAdmin tool, the genus‐ and species‐specific m/z values were revealed. These experimentally detected markers were applied to evaluate the obtained MS data of six model mixtures of insect powders and 10 commercial food products. The insect was identified at the level of genus and species in four model mixtures and in seven food products. At the level of species, the powder model mixtures were identified with 100% efficiency. The identification of insects in food products declaring insect protein content was markedly limited by the presence of other protein sources (chickpea, pea and rice flours), reducing the reliability of the identification. Thus, the markers (genus‐ and species‐specific m/z values) identified in this study can serve as a potential tool for distinguishing among the 10 analysed insect species ( A. domesticus, Alphitobius laevigatus, B. dubia , G. assimilis, H. illucens, L. migratoria, S. gregaria, S. tartara, T. molitor, Z. morio ) in food matrices free from non‐insect proteins. The mentioned markers are more reliable and applicable, especially for quality control of insect powders before their further processing; that is, particularly in the early stages of production of food products with declared edible insect protein content.

Nomenclature

DART‐HRMS

direct analysis in real‐time high‐resolution mass spectrometry

DHB

2,5‐dihydroxybenzoic acid

DNA

deoxyribonucleic acid

HPLC

high performance liquid chromatography

LC–MS/MS

liquid chromatography tandem mass spectrometry

MALDI‐TOF MS

matrix‐assisted laser desorption/ionisation time of flight mass spectrometry

MSP

main spectra profile

SQL

structured query language

TFA

trifluoroacetic acid

Author Contributions

David Straka: investigation, formal analysis, validation, writing ‐ review and editing. Alena Meledina: investigation, formal analysis, validation, writing review and editing, data curation. Tatiana Anatolievna Smirnova: investigation, formal analysis, validation, writing – review and editing. Jana Hajslova: funding acquisition, resources, writing – review and editing. Lenka Kourimska: funding acquisition, resources, writing ‐ reviewand editing. Martin Kulma: resources, writing – review and editing. Katerina Sebelova: resources, writing ‐ review and editing. Ondrej Pospisil: resources, writing – review and editing. Anezka Kopecka: resources, writing ‐ review and editing.  Stepanka Kuckova: conceptualization, formal analysis, funding acquisition, writing – original draft, methodology, project administration, data curation, visualization, supervision.

Funding

This work was supported by the Ministry of Agriculture, Czech Republic, Grant Number: QK23020101; UCT Prague intern grant of Specific university research, Grant Number: A1_FPBT_2025_001; METROFOOD‐CZ, MEYS Grant Number: LM2023064.

Conflicts of Interest

The authors declare no conflicts of interest.

Supporting information

Table S1: Average numbers of peaks found in mass spectra of Tenebrio molitor (killed by cooking) and Gryllus assimilis (frozen to death) obtained under different conditions. (Data were not obtained).

Table S2: Unique m/z values enabling distinguishing insect genus.

Table S3: Unique m/z values enabling distinguishing insect species—crickets.

Table S4: Unique m/z values enabling distinguishing insect species—cockroaches.

Table S5: Unique m/z values enabling distinguishing insect species—darkling beetles.

Table S6: Unique m/z values enabling distinguishing insect species—locusts.

Data Availability Statement

The data that support the findings of this study are available from the corresponding author upon reasonable request.