Aspergillus Metabolome Database for Mass Spectrometry Metabolomics

The Aspergillus Metabolome Database is a free online resource to perform metabolite annotation in mass spectrometry studies devoted to the genus Aspergillus. The database was created by retrieving and curating information on 2811 compounds present in 601 different species and subspecies of the genus Aspergillus. A total of 1514 scientific journals where these metabolites are mentioned were added as meta-information linked to their respective compounds in the database. A web service to query the database based on m/z (mass/charge ratio) searches was added to CEU Mass Mediator; these queries can be performed over the Aspergillus database only, or they can also include a user-selectable set of other general metabolomic databases. This functionality is offered via web applications and via RESTful services. Furthermore, the complete content of the database has been made available in .csv files and as a MySQL database to facilitate its integration into third-party tools. To the best of our knowledge, this is the first database and the first service specifically devoted to Aspergillus metabolite annotation based on m/z searches.


Introduction
The genus Aspergillus is a group comprising both beneficial and pathogenic conidial fungi, which includes more than 300 species present in the environment [1]. Some species produce a wide range of secondary metabolites that are of high industrial and therapeutic importance, such as antibiotics and statins [2]. Aspergillus species usually interact with organisms in an asymptomatic manner and do not cause illness. However, some species have been shown to be responsible for several disorders in organisms with low immunity that have been massively exposed to fungal spores, such as immunocompromised humans [3][4][5][6][7][8], or plants and plant-derived products that have been exposed to mycotoxinproducing molds, leading to food contamination and spoilage [9][10][11]. Invasive aspergillosis can be among the most serious fungal infections [12][13][14][15]. Among the many species of the genus Aspergillus, only some are considered pathogenic. Aspergillus fumigatus is a major cause of disease in humans, followed by Aspergillus flavus, Aspergillus niger, Aspergillus terreus, Aspergillus nidulans, and several members of Aspergillus section Fumigati [13,16,17].
Aspergillosis is a disease with a high mortality rate and it is generally caused by inhalation of Aspergillus fumigatus conidia. These species have a highly adaptable metabolism that allows them to withstand high stress conditions, to face the immune response of the host, to acquire nutrients, or to take competitive advantage during infection. As a result of infection, the production of some fungal metabolites may undergo alterations, reflecting the fungal response to the changes brought by the host. Although aspergillosis has been extensively studied in recent decades, there is still a need for further study of the impact of host-fungal interactions [18].
The most widely used technique to find primary and secondary metabolites, or to perform metabolomic studies, is mass spectrometry (MS) coupled with separation techniques. MS provides the mass-to-charge ratio (m/z) of the ions of interest. However, one of the main challenges in untargeted metabolomics is metabolite identification [37][38][39]. This process typically involves searching for the m/z values in various metabolomic databases, or in a mediator tool that provides a unified interface to those databases [40]. Nowadays there is no available tool to annotate Aspergillus compounds using m/z data obtained by MS instrumentation. There are general databases containing metabolites present in Aspergillus such as Aspergillus Secondary Metabolite Database (A2MDB) [2], KNApSAcK [41], ChEBI [42] or the Dictionary of Natural Products; there are dedicated databases such as FungiDB [43] or AspGD [44], focused on Aspergillus genomics data; and databases such as Aspergillus&Aspergillosis (https://www.Aspergillus.org.uk, accessed on 26 April 2021) devoted to general purposes. NP Atlas [45] is an open-source database containing 24,594 natural products, of which 2029 belong to the genus Aspergillus. However, it only collects the first isolation reference for each compound. Untargeted metabolomic studies are usually trying to annotate compounds and interpret their biological meaning. Therefore, having several references, and having the most up to date and relevant references, is valuable. Moreover, users cannot perform direct searches based on m/z data obtained by means of MS, but they need to perform a manual transformation of the m/z to monoisotopic mass. Although this step is trivial, it requires time, which can be reduced by automatically retrieving the compound candidates for each potential adduct formed. Furthermore, NP Atlas only permits searching for a single mass at a time, which makes the process of searching for all the masses obtained by MS quite tedious. None of the databases mentioned in this paragraph has support for an m/z search for data obtained by MS. Moreover, there are a large number of publications describing metabolites that have been identified in Aspergillus, but the information about many of these metabolites has not been included in any database.
The recent untargeted metabolomic studies on Aspergillus have used internal libraries [30] or general metabolomic databases [29][30][31] for metabolite identification. The first step in performing metabolite identification in MS studies is, in most cases, the annotation process. It consists of the assignment of putative compounds to the m/z obtained by analytical instrumentation as a necessary initial step to inferring biological significance. Currently, this step is often performed by querying general metabolomic databases. However, the number of specific Aspergillus metabolites in databases such as KEGG [46], LipidMaps [47], HMDB [48], or METLIN [49] remains low, which increases the chance that an Aspergillus metabolite is annotated as unknown. Another option is to query general databases with a much higher number of chemical compounds, such as PubChem [50] or ChEBI [42]. However, the high number of hits that are usually obtained in these databases for each m/z makes the subsequent annotation tedious work and increases the likelihood of incorrectly annotating some compounds. The use of internal libraries can be a good solution if the quality of the internal library is high, but it limits the reproducibility of the results and hinders cooperation between research groups. Therefore, the development of a tool specifically devoted to metabolite annotation in Aspergillus studies would provide a great value to the field.
We have integrated information about Aspergillus metabolites present in several metabolomic databases, including KNApSAcK, A2MDB, and ChEBI, and from scientific publications into a single database: the Aspergillus Metabolome Database. It comprises 2811 unique compounds that are present in 601 different species and subspecies of genus Aspergillus. The database also contains meta-information about metabolites, including references to scientific publications where these metabolites are mentioned, and information about the Aspergillus phylogenetic tree. A web tool to query the database based on m/z searches was added to CEU Mass Mediator (CMM) [51], as well as a RESTful API to enable third-party tool integration. The next section presents details on how the compounds were curated and integrated, and how the search tool was constructed. The final result of this process, together with a comparison of the number of Aspergillus metabolites present in the database created in this paper and in the general metabolomic databases, is presented in the results section. Finally, the utility of the service developed is discussed, concluding that it increases the Aspergillus metabolome coverage compared to the main metabolomic databases used in untargeted metabolomics, such as METLIN or KEGG.

Collection and Integration of Aspergillus Metabolites
Different Aspergillus species have been studied for decades and much of this information is scattered in many different resources. An effort was made to collect and organize this information, both from the already available databases, as well as by performing a bibliographic review.
The Natural Product Atlas (NP Atlas), Aspergillus Secondary Metabolite Database (A2MDB), KNApSAcK (a general metabolite database containing information on the species in which these metabolites are present), and metabolites compounds tagged with the Aspergillus role from the general metabolomics database ChEBI were selected for integration because they contain metadata that permits identifying which metabolites from these databases are present in Aspergillus. This metadata allowed us to build an automatic import process for metabolites (all those present in A2MDB, and those present in NP Atlas, KNApSAcK, and ChEBI related to Aspergillus). To this end, custom code was built in the Java language that imported files in JSON (NP Atlas), excel (A2MDB), and HTML (KNApSAcK) format using the GSON (JSON) and the java.io libraries (excel and HTML), respectively. We plan to update the compounds from these sources every 6 months. The compounds were integrated into the CMM database after completing a unification process that is described later in this section. This unification process avoids the presence of duplicates from distinct data sources. The imported information includes the name of the metabolite, its monoisotopic accurate weight, its chemical formula, its three-dimensional structure if available, and the species/species of Aspergillus in which the metabolite was found. When naming the different Aspergillus species, we adopted the nomenclature proposed in [52].
Unfortunately, general metabolic databases such as HMDB, LipidMaps, or KEGG do not contain metadata about the Aspergillus organisms that can be processed automatically by computational tools. Therefore, no information could be extracted from them. Some other databases with interesting information for this project, such as Dictionary of Natural Products or AntiBase, had to be discarded due to their licensing terms.
An extensive literature review has also been carried out to select the articles with information on metabolites present in Aspergillus (Supplementary Information, file "cu-rated_references_Aspergillus_metabolome_db.csv") [53,54]. The purpose of this literature review is twofold: on the one hand, identifying new metabolites not present in the integrated databases to be added manually to the Aspergillus Metabolome Data, and on the other hand, to add meta-information relative to scientific publications where metabo-lites (both those found in the bibliographic review and those imported automatically) are discussed.
To identify metabolites not present in the integrated databases, we faced the challenge of not having well-defined search terms that would permit carrying out a systematic literature search: we are looking for metabolites that are present in Aspergillus, but which we do not know if they have been found in this organism. Therefore, we do not know what to look for. Although we tried to search for generic terms such as "Aspergillus metabolite" or "Aspergillus metabolome", these searches were not effective for our purposes. The literature search started by looking for review papers related to Aspergillus and metabolomics that were published in indexed journals. The metabolites described in them that were not present in any of the integrated databases were incorporated into an Excel spreadsheet, together with their literature references. These references were also further explored to find more metabolites. There are no guarantees that this procedure was comprehensive (it almost certainly was not). However, as we show in the results section, this ad hoc procedure allowed us to find several hundred metabolites that were not present in any of the databases that we have integrated.
Regarding the search for literature references, when in the database from which a metabolite was obtained there was no literature reference associated, the name of the metabolite was searched for together with the word "Aspergillus" to find literature references that could be added to the metabolite's meta-information. In this process, recent papers with a high number of citations were favored, given that they are more likely to provide relevant information to a researcher who has found such a metabolite in a MS study. Both in the search for literature references and in the search for new metabolites, if information about the biological activity of the compound was found in any reference, it was saved in a free text field that was added as meta-information in the Aspergillus Metabolome Database. The literature references and the meta-information are useful for users who can check if the putative metabolites associated with the features obtained by MS means have been previously detected in similar organisms; moreover, users can also consult the references that associate the organism with the putative identification. The complete information about compounds, including literature references and the meta-information, is provided through a RESTful API and in the MySQL backup database.
Sometimes the same metabolite is already present in NP Atlas, A2MDB, KNApSAcK, ChEBI or scientific literature under different names. Due to the existence of several names for the same metabolite, it is also possible that during the bibliographic review a metabolite was included that was already present in some of these databases under some other name. Therefore, a unification process is necessary to eliminate these duplicate entries, and to integrate the meta-information associated with each one of them.
The IUPAC International Chemical Identifier (InChI) was used for performing the unification of compounds. The InChI, in contrast to the authority-assigned identifiers such as CAS, EC Numbers, or CID from PubChem, is derived from the structural formula of the molecule. Therefore, anyone can produce the InChI for a given structure. This identifier is unique: the same InChI always corresponds to the same substance, making it ideal to achieve compound unification through different databases. The InChI can be generated using Mol files (*.mol) containing information about the three-dimensional structure of the molecule. One of the disadvantages of the InChI is that it is variable in length, which can cause complications when storing it in relational databases. The InChI Trust version 1.04 provides a Hash algorithm to generate an InChI Key, which is also unique for each substance but whose length is always 27 characters, making the identifier easier to handle.
NP Atlas, KNApSAcK, and ChEBI compounds contain their InChI. A2MDB compounds do not, neither do most compounds from the bibliographic review. When neither the InChI nor the 3D structure were available, but the name was available, we searched for the name in the CMM database, which contains 186,638 compounds. If the name appeared, and it was manually verified that it was the same compound, the InChI available in the CMM database was used for the compound. If the compound name was not available, or its name did not appear in the CMM database, its structure was drawn using ChemDraw. This tool permits exporting the structure of the molecule in different formats, including InChI. The process of InChI unification prevents the duplication of compounds in the database, provides coherence to all the metadata associated with compounds, and permits the representation of all the isomers as different metabolites (see Supplementary Information, file "SI1_fullERModel.mwb", a MySQL model that can be open using MySQL Workbench to navigate through all the entities present in CMM).
The classification of distinct Aspergillus species in a hierarchy that accurately represents the genus Aspergillus has the potential to permit refining m/z searches to subspecies and provide clues for the biological interpretation of the results. The relation between species and the relation between the compounds, species, and references to scientific articles permits representing useful metadata to the database users (see Supplementary Information, file "SI1_fullERModel.mwb"). Some of the compounds are only related to the general genus Aspergillus and they are not classified into any species, since no studies reporting their identification in specific species have been found. The information about the biological activity of the compounds, when available, has also been included as metainformation in the database. The biological activity is a free text description based on the literature review performed.

Creation of the Search Tool
CMM existing infrastructure was reused for creating the search service for Aspergillus compounds. CMM currently runs on an Apache TomEE 7.1.4 application server and it uses the relational database MySQL Server 8.0.21. CMM entity compound was extended to model the new information regarding Aspergillus compounds. Three new entities were created: Aspergillus compounds, organisms, and references in order to represent and structure the information from each entity respectively (see Figure 1 and Supplementary Information, file "SI1_fullERModel.mwb"). Among other advantages, this information permits categorizing the compounds in different species according to the taxonomy published by Samson et al. [48]. Freemarker templates were used to generate the web browser views to display the entities.
A new web tool was added to CMM to permit querying the databases through the m/z obtained by analytical instrumentation. Optionally, the user may specify the tolerance allowed for the precursor ion and the product ions in Da or ppm (the default value is 10 ppm). The service has all the features already implemented in CMM, including features such as the automatic identification of adducts, or the providing of scores of the likelihood of an annotation based on an expert system [51]. The search can be executed only over the Aspergillus Metabolome Database, or it can also include a user-selectable set of general metabolomic databases (including HMDB, LipidMaps, METLIN, KEGG, FAHFA Lipids, and MINE). The search functionality is also available through a RESTful API to facilitate its integration with third-party tools.
In order to test the web service, 26 Aspergillus metabolites that had previously been identified with reference standards were used (see supplementary material gold_standard_ Aspergillus.csv). The compounds were identified in Aspergillus samples by means of targeted analysis. To collect data about these 26 metabolites, whole culture extracts of Aspergillus fumigatus or Aspergillus nidulans were analyzed by LC-MS and compared with authentic standards. The culture broth containing fungal mycelium was homogenized using an ULTRA-TURRAX (IKA-Werke, Staufen, Germany). Homogenized cultures were extracted twice with a total of 100 mL ethyl acetate, dried with sodium sulfate, and concentrated under reduced pressure. For LC-MS analysis, the dried extracts were dissolved in 1 mL of methanol, while the authentic standards were prepared at a concentration of 1 mg/mL. The samples were loaded onto an ultrahigh-performance liquid chromatography (LC)-MS system consisting of an UltiMate 3000 binary rapid separation liquid chromatograph with photodiode array detector (Thermo Fisher Scientific, Dreieich, Germany) and an LTQ XL linear ion trap mass spectrometer (Thermo Fisher Scientific, Dreieich, Germany) equipped with an electrospray ion source. The extracts or the authentic standards (injection volume of 10 µL) were analyzed on a 150 mm by 4.6 mm Accucore reversed-phase (RP)-MS column with a particle size of 2.6 µm (Thermo Fisher Scientific, Dreieich, Germany) at a flow rate of 1 mL/min, with the following gradient over 21 min: initial 0.1% (v/v) HCOOH-MeCN/0.1% (v/v) HCOOH-H 2 O 0/100, which was increased to 80/20 in 15 min and then to 100/0 in 2 min, held at 100/0 for 2 min, and reversed to 0/100 in 2 min. These 26 metabolites were identified by comparison with reference standards. To test the service, we searched their m/zs in both CMM and METLIN, the general metabolomic database providing an m/z search which is the database containing the largest number of Aspergillus compounds. For comparison, we also searched them one by one in NP Atlas using the monoisotopic masses. A new web tool was added to CMM to permit querying the databases through the m/z obtained by analytical instrumentation. Optionally, the user may specify the tolerance allowed for the precursor ion and the product ions in Da or ppm (the default value is 10 ppm). The service has all the features already implemented in CMM, including features

Results
The Aspergillus Metabolome Database comprises 2029 metabolites from NP Atlas, 614 metabolites from KNApSAcK, 700 compounds from A2MDB that were manually curated, and 491 compounds from the bibliography review, which included 337 papers. After the unification of compounds based on the InChI, the database contains 2811 unique compounds. The complete database is publicly available for download as a MySQL backup, and as three Excel files (https://github.com/albertogilf/ceuMassMediator/tree/master/ CMMAspergillusDB, accessed on 26 April 2021).
The full content of the Aspergillus Metabolome Database, including metadata, was added to the CMM database. This resulted in the addition of 2124 new compounds to the CMM database and the inclusion of metadata for other 203 compounds already present in the database. Three new resources were created in the CMM database to provide support for these compounds. The first resource named "compound" contains information about the compounds (Figure 2a). It includes structural information such as formula, mass, charge type, charge number, InChI, InChI Key, SMILES, and IUPAC Classification; links to other databases such as CAS registry, ChEBI, HMDB, KEGG, KNApSAcK, METLIN, and PubChem; the classification from ClassyFire tool [55]; and the organisms (Aspergillus species and subspecies), ontology terms, and literature references that the compound is linked to. The second resource named "reference" represents the literature references and contains a list of organisms and compounds that are mentioned in the publication (Figure 2b). The last resource represents a specific organism and shows the list of compounds known to be present in the specie and the literature references where these compounds or the species are described (Figure 2c). Information about 2811 compounds associated with 220 Aspergillus species and 381 subspecies, and links to 1514 literature references where these are mentioned, can be consulted through the web resources that were added to CMM (Figure 2a).
This information was included in the MS searches available at CMM, both in the web search tool and in the RESTful API services. A JSON representation of the new entities is available using CMM's API services. A new CMM Card, that is, a web representation of the information of a metabolite, was created to adequately display the new information present in Aspergillus compounds (Figure 2a). The service to annotate Aspergillus metabolites is freely available through the metabolite annotation tool CMM (http://ceumass.eps.uspceu. es/, accessed on 26 April 2021). The source code of the Aspergillus service is available in the CMM Github code repository: https://github.com/albertogilf/ceuMassMediator/ tree/master/ceu-mass-mediator-v4.0, accessed on 26 April 2021. The web application for Aspergillus metabolite annotation is available at http://ceumass.eps.uspceu.es/mediator/ Aspergillus_metabolome_search.xhtml, accessed on 26 April 2021.
The number of metabolites present in the Aspergillus Metabolome Database that are also present in the general metabolomic databases providing batch m/z searches (KEGG, LipidMaps, HMDB, and METLIN) are shown in Table 1, while the coverage between natural products-databases not providing m/z searches is shown in Table 2. In untargeted metabolomics, the annotation of compounds is key and a batch m/z search is usually the first step to annotate, followed by searches in databases containing information about Aspergillus compounds. For untargeted metabolomics researching about Aspergillus the possibility of connecting the two types of databases seems a valuable service. Note that these general databases do not have computer-readable metadata about which metabolites are present in some Aspergillus species. Hence, to construct these tables, we searched for the metabolites present in the databases that also were present in the Aspergillus Metabolome Database using their InChI, or the name of the compound when the InChI was not available in the database. A Venn diagram showing the overlap between Aspergillus metabolites present in the Aspergillus Metabolome Database, KEGG, LipidMaps, HMDB, and METLIN is shown in Figure 3. The monoisotopic masses distribution of the 491 compounds that were not present in other databases is show in Figure 4, and the top 30 classes that they belong to according to ClassyFire are shown in Table 3. This information was included in the MS searches available at CMM, both in the web search tool and in the RESTful API services. A JSON representation of the new entities is available using CMM's API services. A new CMM Card, that is, a web representation of the information of a metabolite, was created to adequately display the new information present in Aspergillus compounds (Figure 2a). The service to annotate Aspergillus metabolites is freely available through the metabolite annotation tool CMM (http://ceumass.eps.uspceu.es/, accessed on 26 April 2021). The source code of the Aspergillus service is available in the CMM Github code repository: https://github.com/albertogilf/ceuMassMediator/tree/master/ceu-mass-mediator-v4.0, accessed on April 26th 2021. The web application for Aspergillus metabolite annotation is available at http://ceumass.eps.uspceu.es/mediator/Aspergillus_metabolome_search.xhtml, accessed on April 26th 2021.
The number of metabolites present in the Aspergillus Metabolome Database that are also present in the general metabolomic databases providing batch m/z searches (KEGG, LipidMaps, HMDB, and METLIN) are shown in Table 1, while the coverage between natural products-databases not providing m/z searches is shown in Table 2. In untargeted metabolomics, the annotation of compounds is key and a batch m/z search is usually the first step to annotate, followed by searches in databases containing information about Aspergillus compounds. For untargeted metabolomics researching about Aspergillus the possibility of connecting the two types of databases seems a valuable service. Note that these general databases do not have computer-readable metadata about which metabolites are present in some Aspergillus species. Hence, to construct these tables, we searched for the metabolites present in the databases that also were present in the Aspergillus Metabolome Database using their InChI, or the name of the compound when the InChI was not available in the database. A Venn diagram showing the overlap between Aspergillus metabolites present in the Aspergillus Metabolome Database, KEGG, LipidMaps, HMDB, and METLIN is shown in Figure 3. The monoisotopic masses distribution of the 491 compounds that were not present in other databases is show in Figure 4, and the top 30 classes that they belong to according to ClassyFire are shown in Table 3.         Regarding the 26 Aspergillus compounds previously identified by comparison with authentic standards, all of them were present in the Aspergillus Metabolome Database, 12 were present in NP Atlas, and 8 were present in METLIN. The results can be seen in the Supplementary Information (gold_standard_Aspergillus_metabolites.csv).

Discussion and Conclusions
The dispersion in the distinct data sources is a general issue in metabolomics. In the case of Aspergillus, this dispersion was compounded with a large number of references generating knowledge about Aspergillus which were not collected in metabolomic databases. Both these situations hinder the annotation process and slow down research on Aspergillus.
The work presented here tackles this problem by creating the Aspergillus Metabolome Database. This database was created by compiling metabolites present in Aspergillus from public databases whose license allowed their integration into a third-party database (NP Atlas, KNApSAcK, A2MDB, and ChEBI), as well as through a literature review that included 337 articles. These compounds were unified by means of their InChI to avoid duplications. After the unification process, the database contains 2811 unique Aspergillus metabolites and 1514 papers with information on those metabolites. In addition to the name of the compound, its formula, and its m/z, the database contains meta-information, such as taxonomic information on the different Aspergillus species, and references to publications related to the compound.
A comparative analysis of the number of Aspergillus metabolites contained by the general metabolomic databases was performed. The general database that contained the highest number of Aspergillus compounds present in the Aspergillus Metabolome Database was METLIN, which contained 214 compounds. This means that approximately 92% of the compounds available in Aspergillus Metabolome Database were not present in METLIN. Even if we consider all general metabolomic databases together, 2535/2811 (90%) of the compounds present in the Aspergillus Metabolome Database were not present in any of them (see Table 1 and Figure 3). This suggests that if the general metabolomic databases are used to annotate Aspergillus compounds, it is highly likely that an m/z corresponding to a known compound returns no hits, and results in an annotation of the compound as unknown, even when searching for the m/z in all of them.
Due to the lack of metadata regarding which compounds are present in some Aspergillus species in the general metabolomic databases, it is not possible to know if there are compounds identified in Aspergillus species present in them that have not been included in the Aspergillus Metabolome Database. This also means that, if such compounds existed, it is not possible to include them in the Aspergillus Metabolome Database. However, the fact that most of the compounds present in the Aspergillus Metabolome Database are not present in these databases (see Table 1 and Figure 3) leads us to believe that if such compounds exist, they should be very few.
A web tool was built and integrated into CMM that permits searching for m/z over the Aspergillus Metabolome Database, specifying the tolerance of the mass, the possible adducts formed, the chemical alphabet, the modifiers used, and the inclusion or exclusion of general databases. This functionality is also available through a RESTful API, which can permit its integration of m/z searches of Aspergillus compounds in third-party tools. The source code for the web search tool and for the RESTful API has been released under an open license. In addition, all the current 2811 metabolites of the Aspergillus Metabolome Database have been published as Excel files and as MySQL backup files.
However, new Aspergillus metabolites are continually being discovered. We plan to continue updating the Aspergillus Metabolome Database with this information in the future, and we strongly encourage the scientific community to contribute information from internal databases on Aspergillus metabolites to this project. We believe that having a unified open repository for all Aspergillus metabolites, as well as tools that permit querying it, is a benefit for the entire scientific community involved in studying Aspergillus species.
Supplementary Materials: The following files are available at the link https://github.com/albertogilf/ ceuMassMediator/tree/master/CMMAspergillusDB. File "SI1_fullERModel.mwb": MySQL workbench full entity relation model of the CMM information system. File "curated_references_Aspergillus_ metabolome_db.csv": list of references and sources reviewed to populate the database. File "com-pounds_Aspergillus_metabolome_db.csv": list of compounds included in the Aspergillus Metabolome Database. File "organisms_Aspergillus_metabolome_db.csv": list of Aspergillus species and subspecies included in the Aspergillus Metabolome Database. File "references_Aspergillus_metabolome_ db.csv": list of full references linked to Aspergillus compounds in the Aspergillus Metabolome Database.

Conflicts of Interest:
The authors declare no conflict of interest.