Crops are exposed to abiotic and biotic factors that can negatively impact their productivity, i.e., floods, droughts, degraded soils, and phytopathogens. Phytopathogens are responsible for up to 40 % of crop yield reductions (^{Valenzuela-Ruiz et al., 2020}). Along with this, the growing demand for food fueled by population increases has led to the modification of agricultural practices to mitigate the losses caused by phytopathogens, without lowering the nutritional quality and safety of the products grown (^{FAO, 2018}). A widely used strategy to address these aspects has been the use of synthetic pesticides, which increased 97.7% worldwide from 1992 (1.30 kg ha^-1) to 2016 (2.57 kg ha^-1) (^{FAO, 2018}); in Mexico, the use of synthetic fungicides increased 76.9% (from 1.08 to 1.91 kg ha^-1) in the same period (^{Roser, 2019}). Reports show that only 0.1% of the applied synthetic pesticides reach the crop of interest (^{Gill and Garg, 2014}), a fact that creates serious economic and environmental impacts, such as soil and aquifer contamination with recalcitrant wastes, soil, chemical, and microbial degradation, and biodiversity loss (^{de los Santos-Villalobos et al., 2018}).

In this context, phytopathogen management in cropping systems must be extensively and comprehensibly addressed to be consistent with Goal 2, end hunger, of the 2030 Agenda for the Sustainable Development for Latin America and the Caribbean. Goal 2 establishes the need to concentrate efforts to achieve food security, improve nutrition, and promote sustainable agriculture (ONU, 2018). Mexico has collaterally contributed to these goals since 1940 when the first studies and applications of biological control began and academic programs were established to promote an integrated plant health management (Gutiérrez-Samperio, J. Personal communication. 2019). Biological management of phytopathogens is defined as “the use of beneficial organisms to reduce the negative effects caused by phytopathogens, through their antagonistic actions, either by direct actions of gene recognition or indirectly through metabolites or products derived from them” (^{Valenzuela-Ruiz et al., 2020}). In a broader sense, biological management was initially promoted through campaigns implemented by government agencies, where predator insects, parasitoids or sterile insects were used, followed by the inclusion of entomopathogenic fungi (Beauveria bassiana and Metarhizium), filamentous fungi and bacteria, to control important pests nationally (^{Arredondo-Bernal and Rodríguez-del Bosque, 2015}).

The use of biological control agents of microbial origin (BCA-M) in Mexico is an alternative to complement the conventional control of phytopathogens using synthetic pesticides (^{Arredondo-Bernal and Rodríguez-del Bosque, 2015}). For example, in Mexico, the area sown in 2017 was approximately 32.4 million hectares, where the main crops were maize, sorghum, bean, coffee, sugarcane, and wheat (^{INEGI, 2017}); the main phytopathogen genera reported in those crops were Fusarium, Colletotrichum, Leveillula, Botrytis, Rhizoctonia, Pythium, Phytophthora, Pseudomonas, Candidatus Liberibacter asiaticus, Leifsonia, and Xanthomonas, and the BCA-M that were most used for controlling these pathogens were the fungal species Trichoderma harzianum, Paecilomyces lilacinus, P. fumosoroseus and Beauveria bassiana, and the bacterial species Bacillus subtilis, B. pumilus, B. amyloliquefaciens, B. licheniformis and B. thuringiensis (^{García-Juárez et al., 2016}; ^{Villarreal-Delgado et al., 2017}; ^{Delgado-Ortiz et al., 2019}). Although researches on BCA are significant (Tables 1 and 2), sometimes, no detailed information on the results of their application to nationally important crops in the field are published or are included in technical reports with restricted access and limited to describing crop yield increases and/or phytopathogen inhibition.

Identifying new agents and/or understanding the plant-pathogen-BCA of microbial origin interaction (under field conditions), as well as conducting robust studies of the polyphasic taxonomic affiliation (consensus classification based on the integration of all available phenotypic and genotypic data), action and ecology mechanisms of those BCA in the agroecosystems, using the recent advances in omics sciences, are essential for registration, marketing, innovation and success of formulated biopesticides. For this reason, in the framework of the international celebration of Plant Health, the objective of this review is to describe the knowledge status and analyze the aspects which limit the bioprospection and extensive use of BCA-M in Mexico, cutting-edge genomic strategies inherent to the omics sciences focused on the identification, selection and study of the mechanisms of action of those agents, and the dissemination and socialization of the scientific knowledge gained to enhance agro-biotechnological innovation and biopesticides success.

Biological control agents of microbial origin (BCA-M) and biopesticides in Mexico. The identification and introduction of natural enemies in economically important phytopathogens in Mexico became relevant in the 1940s due to the need to develop efficient and sustainable management strategies. In the more than 80 years of using BCA-M in our country (^{Bernal and Quezada, 1999}), 230 bioproducts from 40 companies registered by COFEPRIS (https://www.gob.mx/cofepris) are available in the national market. However, some of these products have a limited number of active ingredients at the taxonomic group level (fungi, bacteria y/o viruses), or microbial genus/species for controlling the causal agents of diseases in plants that are cultivated in Mexico (Table 1). This reflects a limited availability of bioproducts to be used in the Mexican fields despite the efforts made and the research conducted in our country. This fact could be associated with the limited support to the scientific sector to conduct systemic research, bioprospecting new BCA-M, and obtaining patent registration of biotechnological products, as well as with the limited organizational structure at the institutional level to establish departments specialized in patent and licensing processes. The lack of financial resources has often impacted the development of partial research, mainly in vitro, from collections with limited biological and regional representation. On the other hand, there is no access to documentary information to scientifically support their effectiveness, nor statistical data of their production, consumption and success at the national level (^{De León and Mier, 2010}), with the exception of products such as Fungifree^®, which is an effective biofungicide for managing foliar diseases such as anthracnose (C. gloeosporioides, C. acutatum and C. fragariae), powdery mildew (L. taurica, E. chichoracearum and S. humili) and gray mold (B. cinerea) (^{Galindo et al., 2015}), and the studies conducted by Centro Nacional de Referencia de Control Biológico SAGARPA-DGSV for controlling Diaphorina citri and lobster with entomopathogenic fungi (^{Ayala-Zermeño et al., 2015}; ^{Gallou et al., 2016}).

Furthermore, one of the main constraints for technology transfer and extensive use of bioproducts based on BCA-M is the insufficient publication and dissemination of scientific research conducted by different universities and research centers in the country, even after the product’s patent has been obtained (^{Bernal and Quezada, 1999}). In this regard, the important role of the Mexican Journal of Phytopathology or RMF, for its acronym in Spanish (https://www.rmf.smf.org.mx/), must be recognized, because during the past eight years, different researchers have published about 50 scientific papers (23.7% of the total number of the published articles) reporting aspects related to the use of BCA-M, i.e., scientific articles, review articles, phytopathological notes and phytopathological reports (Table 2). Likewise, records provided by Scielo Analytics (https://analytics.scielo.org/) show that of the 100 articles of the JMP that are most consulted online, 24% correspond to subjects related to BCA-M, of which the following stand out: the Bacillus genus as a biological control agent and its implications in agricultural biosafety (^{Villarreal-Delgado et al., 2017}), the action mode of Candida oleophila for managing Penicillium expansum and Botrytis cinerea (^{Guerrero et al., 2011}), and “witch broom” biocontrol with Trichoderma spp. in mango crops under field conditions (^{Michel-Aceves et al., 2013}), just to name a few examples.

In this regard, to strengthen the scientific research and enrich the functional range of bioproducts registered and sold in Mexico, it is essential to incorporate new cutting-edge and robust tools into scientific studies of BCA-M, especially about their polyphasic taxonomic affiliation, action mode in vitro, and under greenhouse and field conditions, as well as their interaction with biotic and abiotic factors. This would lead to the establishment of cost-effective production, formulation, scaling and innovation systems (^{Galindo et al., 2013}; ^{Serrano-Carreón et al., 2010}), which, in turn, would help governmental agencies and professionals involved in agricultural production make an efficient, safe, and rigorous technology transfer of these bioproducts to farmers.

Taxonomic affiliation of BCA-M. The molecular taxonomic study of BCA-M has been conventionally undertaken using the 16S rRNA gen for prokaryote taxonomic affiliation, because i) it is present in that group of organisms and often exists as a multigenic family; ii) its function has not changed over time, thereby suggesting that slow changes occur in its sequence during the evolutionary course; and iii) its length (1500 bp) is great enough for technical and bioinformatic purposes but it is difficult to identify bacterial species because of its limited genetic diversity (^{Aguilar-Marcelino et al., 2020}). Likewise, the ITS region (internal transcribed spacer) in eukaryotes, which is made up of the ITS1, 5.8S rRNA gene and ITS2, is highly conserved at the genus level, whose length (650-1,100 bp) and limited genetic diversity among species hampers their correct differentiation and taxonomic affiliation. The BCA-M taxonomic affiliation is based on the use of molecular techniques (RAPDs, ISSR, RFLP, AFLP) and/or PCR for rapid detection using marker genes, such as 16S rRNA, 28S rRNA, ITS, β-tubulin, RPB2, elongation factor, SCAR (^{Mazrou et al., 2020}; ^{Hassan et al., 2019}). The taxonomic resolution at the species or subspecies level (i.e., haplotypes) will depend on the preferred molecular technique, restriction enzymes, and selected genes or regions.

The correct taxonomic affiliation is decisive for bioprospecting biosafe BCA-M, considering that several genera have species with antagonistic activity against phytopathogens but may also have other species that cause diseases in plants, animals and/or humans. For example, Bacillus is one of the bacterial genera that has been most reported and studied worldwide because of its biological control capacity. The B. cereus group has been extensively studied and is made up of, at least, eight species that are closely related: B. anthracis, B. cereus, B. thuringiensis, B. mycoides, B. pseudomycoides, B. weihenstephanensis, B. cytotoxicus and B. toyonensis. However, the members of this group have a significant impact on human health, agriculture, and the food industry, i.e., B. anthracis is the etiological agent of anthrax, an obligated pathogen that poses a threat to human and herbivores health, while B. thuringiensis is a biological management agent which is extensively used and recognized as a safe biopesticide in the world (^{Liu et al., 2015}). On the other hand, B. cereus is an opportunistic pathogen that often causes food poisoning and has recently been reported as an emerging phytopathogen. At the intraspecific level, members of the B. cereus group have sequences of the 16S rRNA gene that are very similar and highly conserved genomes (^{Ehling-Schulz et al., 2019}), a fact that makes their correct taxonomic affiliation difficult. Another group of taxonomic importance with a molecular approach is the Burkholderia genus, due to biochemical, biological, and morphological characterization restrictions (^{Espinosa-Victoria et al., 2020}; ^{Serret-López et al., 2021}; ^{de los Santos-Villalobos et al., 2018}). In recent years, the introduction of various next-generation sequencing strategies and platforms (NGS), as well as genomic, phylogenomic, and bioinformatic tools, have significantly contributed to a better understanding of the microorganism’s evolution and interaction with their environment (^{Jagadeesan et al., 2019}). These tools have also made it possible to study microbial strains at the genomic level, which has strengthened their correct taxonomic affiliation and identification of the genes involved in colonization and adaptation to the host, and their mechanisms of action, among other aspects (Figure 1).

Genomic studies for the taxonomic affiliation of BCA of microbial origin (BCA-M). At present, the robust taxonomic affiliation of microorganisms is based on the extraction, sequencing, and analysis of their genome, partial or total, supported by morphological, physiological, and metabolic traits (Figure 2). For this, sequencing high-quality DNA of the microorganism of interest and validating the obtained data through a quality control process are decisive factors, because reads of DNA without the required quality can be excluded in subsequent analysis (^{Xi et al., 2019}). FastQC is a digital tool designed to perform numerous quality control checks of the sequences that are obtained (^{Andrews, 2010}). Then, the adapters must be trimmed, and low-quality data of the obtained genomic information removed using programs such as Trimmomatic v0.36 (^{Bolger et al., 2014}) and Cutadapt v1.18 (^{Martin, 2011}). The genome is assembled using programs such as SPAdes v3.14.1, which assembles non-ribosomal peptides associated with biological control mechanisms (^{Bankevich et al., 2012}), hybrid SPAdes (^{Antipov et al., 2016}), Velvet v1.2.10 (^{Zerbino, 2008}), IDBA-UD v1.1.1 (^{Peng et al., 2012}), Canu (^{Koren et al., 2017}) and SOAPdenovo2 (^{Luo et al., 2012}). Once the genome has been assembled, the obtained contigs are rearranged based on a reference genome with the Mauve contig Mover 2.4.0 program (^{Darling et al., 2004}). Finally, the genome annotation is carried out to identify the encoding sequences and associate them with a function or identify hypothetical proteins. For this, the National Center for Biotechnology Information (NCBI) provides an automatic prokaryotic genome annotation tool. Similarly, RAST and xBASE2 are web servers used to annotate genomes of bacteria and archaea (^{Aziz et al., 2008}; ^{Chaudhuri et al., 2008}). Genome contamination can be quantified using different software packages, such as checkM and Quast, for prokaryotes and eukaryotes (^{Gurevich et al., 2013}; ^{Parks et al., 2015}). Recently, bioinformatics tools in silico have been developed for taxonomic affiliation of bacterial species (Figure 2), of which the most important is the overall genome relatedness indices (OGRI), such as the average nucleotide identity (ANI) and the genome-to-genome distance calculator (GGDC), to replace the DNA-DNA hybridization standard (DDH) (^{Richter et al., 2016}). This has been used to study prokaryotes only since there is no equivalent OGRIs for fungal strains (^{Raja et al., 2017}).

ANI is widely used to establish similarity between two prokaryote species at the genomic level (^{Yoon et al., 2017}) and is defined as the average percent of nucleotide sequences identity of orthologous genes shared between two genomes (^{Sentausa and Fournier, 2013}). The ANI value of the genome under study corresponds to the average of the identity values among all the pairs of orthologous fragments between two genomes. The ANI value used to affiliate a strain to a bacterial species previously described is ≥ 95-96% (^{Varghese et al., 2015}; ^{Moriuchi et al., 2019}).

Figure 1 A. Mechanisms of action that have been most studied and reported in biological control agents of microbial origin (BCA-M) (Source: Authors). B. Multiple biotic-abiotic functions that can be associated with an BCA-M, Burkholderia case (Source: Espinosa-Victoria et al., 2020). 

Figure 2 Workflow of the genomic and metabolic methodological processes used in taxonomic research and to establish the mechanisms of action of biological control agents of microbial origin (BCA-M) to suppress phytopathogens. 

On the other hand, unlike ANI (index similarity type), GGDC (http://ggdc.dsmz.de/) is an OGRI based on the distance ratio which enables the alignment of the genome of interest with one of reference; then a set of high-scoring segment pairs (HSP) is obtained, known as “intergenomic coincidences”, from which specific distance formulae are calculated to transform the HSP data to a GGDC value (^{Meier-Kolthoff et al., 2013}), using a >70% GGDC value to taxonomically affiliate a strain to a bacterial species previously reported.

Currently, the use of tools and strategies focused on robust and comprehensive BCA-M taxonomic affiliation (based on genome sequencing and polyphasic studies) is experiencing a boom in Mexico. For example, one of these studies was recently published by ^{de los Santos-Villalobos et al. (2019)}, who reported a new bacterial species with biological control capacity against Bipolaris sorokiniana, an emerging agent that causes spot blotch in wheat in the Yaqui Valley, Mexico. Initially, this bacterial strain (TE3^T) was affiliated to the Bacillus genus based on the partial sequencing of the 16S rRNA gene (Valenzuela-Aragón et al., 2018). Then, ^{Villa-Rodriguez et al. (2019}) proposed its taxonomic affiliation to the B. subtilis species by sequencing the whole 16S rRNA gene. Finally, the sequencing and study of the whole genome, the OGRIs analysis (ANI and GGDC), and the polyphasic characterization of strain TE3^T allowed a robust taxonomic affiliation of this BCA to a new bacterial species, termed B. cabrialesii (^{de los Santos-Villalobos et al., 2019}). Bacillus cabrialesii TE3^T inoculation in wheat plants infected by B. sorokiniana mitigated the disease by reducing the 3-5 severity scale of leaf infection and the density of lesions to 3.06 ± 0.6 cm^-2 leaves compared to plants inoculated with the phytopathogen (8-9 and 6.46 ± 1.46 lesions cm^-2).

Study of the action mechanisms in BCA-M. In recent decades, studies of the effect of BCA-M in phytopathogens management have been carried out. This has been useful to elucidate the different mechanisms used by these agents to reduce the growth, development, and infection of phytopathogens in cultivated plants. Among the interaction mechanisms with suppressive effects that have been most studied and reported are i) mycoparasitism; ii) antibiosis; iii) lytic enzymes activity; iv) siderophore biosynthesis; v) production of δ-endotoxins; vi) lipopeptides biosynthesis; and vii) induced systemic response (Figure 1A). However, since the action of the BCA-M involves non-parasitic phases that are essential for their adaptation to the edaphic or endophytic environment and ensure their suppressive activity, they must be systematically studied to design soil, substrate, or plant bioremediation strategies (Figure 1B).

In Mexico, based on the contributions published by the Mexican Journal of Phytopathology and other national and international scientific journals, there are many scientific studies focused on the selection and evaluation in vitro of promising BCA-M. Although to a lesser extent, the specific mechanisms of action previously mentioned have been also explored (Table 2 and Figure 1). However, the innovation of biopesticides developed in Mexico and other countries requires that new BCA along with the mechanisms responsible for the activity observed be identified, both in new strains and in those that have been already reported, because the mechanisms or metabolites used by those BCA are not necessarily responsible for the observed effect or are directly related to their bioactivity in the field. Consequently, state-of-the-art and integrating strategies that allow a better understanding of the mechanisms of action of the BCA-M under field conditions must be implemented. This would increase the probabilities of success in implementing these agents in the agricultural sector, thereby optimizing the production processes, scaling new bioproducts, and promoting the innovation of those already registered and available on the market.

Genomic studies of BCA-M to identify potential mechanisms of action. So far, several strategies and platforms have been developed to identify clusters of genes in genomes associated with biological control, especially those related to non-ribosomal peptide synthetases (NRPS) and polyketide synthases (PKS) (Figure 2). These secondary metabolites exhibit a significant variety of biological activities, and many of them are clinically valuable antimicrobial, antifungal, antiparasitic, antitumor, and immunosuppressant compounds (^{Ansari et al., 2004}). Among the tools to identify those clusters of genes in the genome of BCA of bacterial origin are PRISM (^{Skinnider et al., 2017}), CLUSEAN (^{Weber et al., 2009}), RIPPMiner (^{Agrawal et al., 2017}), and for fungal genomes, SNAP (^{Jhonson et al., 2008}), Augustus (^{Keller et al., 2011}) and GeneMark-ES (^{Borodovsky and Lomsadze, 2011}). However, the AntiSMASH 5.0 and BAGEL4 tools have been widely reported and used for these purposes, but the identification of genes associated with the BCA-M activity using these tools is predictive, and, for this reason, their expression and function in vivo must be further studied and complemented with other strategies that will be later described.

AntiSMASH 5.0 uses the HMMer3 tool (http://hmmer.janelia.org/) and the translations of the amino acid sequence of all the encoding genes, which are analyzed with profile hidden Markov models (pHMM) based on alignments of multiple sequences associated with experimentally characterized proteins or proteins/domains (^{Blin et al., 2019}). The detection stages use a filtering logic of negative and positive pHMM and their cut-off points, based on the knowledge of the minimum basic components of each type of clusters of the genes reported in the scientific literature.

On the other hand, BAGEL4 translates into six proteins [one for each open reading framework (ORF) Α, each of the encoding genes of the BCA genome that is being studied (^{Van Heel et al., 2018}). These proteins are examined to detect the occurrence of certain protein motifs (elements conserved in an amino acid sequence that is usually associated with a specific function) and compared to a peptide database. Finally, the areas of interest (AOI) are identified and thoroughly analyzed. The ORF in AOI are first analyzed using Glimmer3 (^{Delcher et al., 2007}). This program is set up in such a way that Glimmer3 creates a model for each defined AOI. Later, small ORF are named in the intergenic regions. The default setup for these ORFs is a minimum length of 72 bp (24 amino acid residues) and a superposition of 10 bp is allowed with the ORF of Glimmer3.

Currently, the use of these tools for the integrative approach of the potential mechanisms of action of BCA-M are being exponentially developed in Mexico. An example of this is the research of ^{Valenzuela-Ruiz et al. (2019)}, who used this strategy to identify an BCA-M taxonomically affiliated, by studying its genome and OGRIs, as Bacillus paralicheniformis TRQ65. The study of the clusters of genes associated with NRPS and PKS of the B. paralicheniformis TRQ65 genome made it possible to identify nine genes associated with the lipopetide biosythesis: lichenicidin, haloduracin, bacteriocin, enterocin and sonorensin (by BAGEL4), and eight genes associated with the lipopeptide biosynthesis: bacillibactin, butirosin, lichenysin, bacitracin, and haloduracin (by antiSMASH). The functionality of those genes was validated through essays of dual confrontation against Bipolaris sorokiniana, where an inhibition area of 1.6 ± 0.4 cm was observed, which confirmed the functionality of the putative genes found in the genome of strain TRQ65, associated with biological control of phytopathogens (^{Villa-Rodríguez et al., 2019}; Valenzuela-Ruiz et al., 2019).

Transcriptomic studies to identify potential mechanisms of action of BCA-M. The modulation in the gene expression of an organism, associated with its interaction with biotic and/or abiotic factors, is found by conducting transcriptomic studies, which are focused on the study of the gene expression, co-expression patterns, signaling cascades, and regulated molecular mechanisms during a biological condition and specific time (^{Aguilar-Marcelino et al., 2020}).

The transcriptomic studies are highly dependent on high-quality total RNA extracted under the specific conditions of the study, followed by sample preparation and sequencing (Figure 2). The reas taken are filtered through a quality analysis using FastQC (^{Andrews, 2010}), the adapters and low-quality reads are removed using Trimmomatic v0.36 (^{Bolger et al., 2014}) or Prinseq (^{Schmieder and Edwards, 2011}). Subsequently, the reads are aligned with and mapped to a reference genome (or transcriptome) to identify and quantify the regulation of the genes and find out what their allelic variants are (^{Rollano-Peñaloza and Mollinedo-Portugal 2017}). The organisms whose genomes have no introns are processed with contiguous aligners, such as Bowtie2 (^{Langmead and Salzberg 2012}) or BWA (^{Li and Durbin 2009}), which are designed to align DNA. Conversely, the best option when there are genomes with introns is the use of aligners such as Tophat2 (^{Kim et al., 2013}), Hisat2 (Kim et al., 2013), STAR (^{Dobin et al., 2013}), GSNAP (^{Wu and Nacu 2010}), SOAP2 (Li et al., 2009) or Kallisto (^{Bray et al., 2016}). The assembly of transcripts and estimation of abundances is usually performed with Cufflinks and Cuffmerge (^{Trapnell et al., 2009}), but there are also reports of transcriptomic studies using DESeq (^{Love et al., 2014}) or HTSeq-count (^{Anders and Huber, 2015}).

The aligned sequences can be arranged with Samtools (^{Li, 2011}) and Picard (http://broadinstitute.github.io/picard/). This information can be used by programs such as Bedtools (^{Quinland and Hall, 2010}) to detect single-nucleotide polymorphisms (SNP) (^{Rollano-Peñaloza and Mollinedo-Portugal, 2017}). The appropriate analysis of the gene expression is based on a standardization due to the different genome sizes and depth variation in each sample sequencing. Therefore, when studying de novo transcriptomes, it is recommended to use RSEM software (^{Li and Dewey, 2011}) because it also provides standardization values. Then, to identify the differentially expressed genes, a genomic annotation must be performed by aligning the sequences of the genes that were identified as significantly expressed (^{Camacho et al., 2009}). Then, the gene families and/or their products are identified using software packages such as HMMER (^{Potter et al., 2018}), InterProScan (^{Mitchell et al., 2015}), and SignalP (^{Petersen et al., 2011}). Finally, a gene ontology analysis (GO) is carried out, whereby genes are grouped by biological processes, molecular function, cell component and metabolic pathway analysis (KEGG) (^{Kanehisa et al., 2004}) and other more complex groups (PANTHER) (^{Mi et al., 2013}), where tools such as AgriGO (^{Tian et al., 2017}) are used to visualize, through hierarchical trees, the GO terms that are significantly expressed. Another approach can be the use of primers or primers of specific genes that have previously shown mechanisms of action of interest to quantify the gene expression through qPCR or ddPCR. This technique avoids the transcriptomic study of an organism.

Several reports have been published worldwide about the use of transcriptomics focused on understanding the biological control mechanisms during the plant-pathogen-BCA interaction by conducting studies of the gene expression in BCA-M (in the presence of a phytopathogen in vitro or in the field); the regulation of the gene expression in the pathogen or in the host plant (in the presence of the BCA-M or its metabolites); the impact of the phytopathogens on the gene expression profiles; and the modulation of these profiles by the presence or absence of an BCA-M. For example, ^{Duke et al. (2017}) reported that Brassica napus (canola) is greatly affected by Sclerotinia sclerotiorum, the causal agent of canola stem rot. However, when Pseudomonas chlororaphis PA23 was inoculated in canola in the presence of S. sclerotiorum, the infection rate was reduced by 91%. This result was associated with the transcriptional reprogramming in canola, i.e., the plants infected with S. sclerotiorum positively regulated approximately 8000 genes, while the preventive treatment with P. chlororaphis PA23 reduced 16 times the expression of the genes previously regulated by the phytopathogen. In addition, in the absence of the pathogen, P. chlororaphis PA23 caused the regulation of defense genes in canola, which are involved in the induced systemic response and production of reactive oxygen species.

Metabolomic studies to identify potential mechanisms of action of BCA-M. As a scientific approach based on data to detect and quantify hundreds of metabolites by differential analyses (^{Lloyd et al., 2015}), metabolomics is ideal for analyzing complex interactions at the metabolic level during plant-microorganisms-environment interactions.

The metabolomic studies are based, in the first place, on establishing the experimental condition for the study (Figure 2) by which the strategy of sample preparation will be established (focused on the type of compounds to be detected, including extractions with organic solvents or solid-phase extractions), followed by sample preparation, either by concentration or purification (^{Mhlongo et al., 2018}). Then, the separation and detection of metabolites of interest is carried out using gas chromatography (GC), liquid chromatography (LC) or capillary electrophoresis (CE) coupled to mass spectrometry (MS) (^{Naz, 2014}). CE separates the compounds depending on the charge and size and provides high-resolution power. CE-MS is mainly used for primary-intermediate metabolic pathways and is usually coupled to a TOF mass analyzer (^{Ramautar et al., 2016}). In recent years, MS imaging (MSI) has evolved and has been used in different metabolic studies to obtain the distribution of compounds in tissue areas (cells, tissue, or specific sections) (Mhlongo et al., 2018). Compared to other traditional molecular imaging techniques, MSI makes it possible to obtain more information by providing the distribution of characteristics of a wide range of metabolites at high resolution (^{Schwamborn, 2012}).

Metabolomics, the same as other omics tools, produces a large amount of complex data which require storage and processing instruments. These data can be processed using free statistical tools such as MarVis1, Mzine, XCMS, MAVEN, Metaboanalyst and MetAlign (^{Benton et al., 2008}), or commercial programs: Markerlynx (Waters), profile creation solutions (Shimadzu), Mass Profiler.pro (Agilent) and metabolic profiler (Bruker) (^{Mhlongo et al., 2018}). Multivariate statistical tools are also used to reduce data dimensionality, variable discrimination, and to form groups based on shared characteristics among samples (i.e., n-strains); part of these are the Principal Component Analysis and Orthogonal Projections to Latent Structures Discriminant Analysis (OPLS-DA) (^{Worley and Powers, 2013}). Finally, the metabolites are annotated and identified using complementary databases which contain mass ranges, compound name and structures, statistical models, and metabolic pathways (^{Fukushima and Kusano, 2013}). Recently, several databases have been developed that contain statistical and metabolomic tools based on MS or nuclear magnetic resonance, i.e., MeRy-B, MeltDB, and SetupX (Fukushima and Kusano, 2013).

The addition of metabolomic approaches to understanding the mechanisms of action of the BCA-M are highly valuable for the agricultural sector due to the wide diversity of metabolites produced by the BCA-M with high agro-biotechnological potential, and the knowledge of the metabolites involved in phytopathogens biological control during the BCA-host-phytopathogen-environment interaction. For example, ^{Vinale et. al. (2014}), using metabolomic approaches, identified a new metabolite produced by Trichoderma harzianum, called isoharzianic acid, which inhibits the growth of phytopathogens such as Sclerotinia sclerotiorum and Rhizoctonia solani. The isoharzianic acid also improves tomato seed germination and induces systemic resistance in plants. Similarly, other metabolites associated with biological control were reported for Trichoderma, such as hetelidic acid, sorbicilinol, trichodermanone C, giocladic acid and bisorbicilinol (^{Kang et al., 2011}).

It is essential that in Mexico important agricultural producer support be strengthened for cutting-edge scientific research focused on the polyphasic taxonomic affiliation of BCA-M and the integrative study of the knowledge of the mechanisms of action of those agents during the interaction with the host plant, the phytopathogen, and the environment. This will strengthen the development, innovation, and success of the use of formulated and registered bioproducts and thus provide a sustainable and ecological alternative to benefit crop health.

Agroecological perspectives of BCA-M supported by omics sciences. The use of BCA-M is a functional and economically viable alternative for controlling phytopathogens. However, it is important to create and strengthen regional stocks to identify potential BCA-M and delve into the study of their mechanisms of action under field conditions, since this will make it possible to know with a greater degree of certainty what the diverse aspects that impact their establishment and functionality are, such as specific suppression actions, as well as the action threshold density, population dynamics, interspecific competence and mechanisms of physicochemical adaptation to the edaphic environment and plant. This knowledge would help establish doses, frequency, and multiplication-application substrates.

Among the aspects identified that are determinant to increase the efficiency of the BCA-M actions in the field, is the activation of the induced systemic response (ISR) through the recognition of microorganisms-associated molecular patterns (MAMPs) to pathogens (PAMPs), or damage (DAMPs) (Figure 2). This kind of signal in plants makes it possible to detect the pathogen and reduce its colonization in the infection spot, and even activate the systemic mechanisms in the plant’s distal tissues (^{Burketova et al., 2015}). Thus, ISR is a plant status which, through biotic factors (i.e., BCA-M) or chemicals, protects itself against continuous and future phytopathogen attacks (^{Ku´c, 1982}). This process was first reported when the phytopathogen came into contact with the root, but studies about the colonization of resistance inductors through foliar applications have proposed the initial mechanisms of that resistance. For example, ^{Lamdan et al. (2015}) reported that the BCA belonging to the Gliocladium virens and G. atroviride species interact with the plant and the pathogen, thereby promoting the plant’s growth and inducing a systemic response through proteins that are secreted by these fungi, which reduces the disease in the aerial parts of the plant. In this way, microbial elicitors induce the systemic response and act as endogenous or exogenous activators by reprogramming the resistance genes (^{Gupta and Bar, 2020}). Ergosterol, the main component of fungal cell membranes and action substrate of many chemical fungicides, has acted as a resistance inductor in tomato and tobacco (^{Granado et al., 1995}; ^{Vatsa et al., 2011}). On the other hand, host and non-host plants of Cladosporium fulvum (tomato pathogen) showed that the CfHNNI1 gene induces a hypersensitive response in tomato and tobacco (^{Xu et al., 2012}), which coincides with the report of ^{Zhang et al. (2013}), who mentioned that the Sclerotinia sclerotiorum protein elicitor SCFE1 induced immunity through PAMP in Arabidopsis thaliana. Nevertheless, compared to bacteria and oomycetes elicitors, the fungal elicitors are not yet totally understood, and the function of those detected so far is still unknown, so the need to conduct further studies prevails (^{Liu et al., 2013}).

In this way, healthy plants can be treated with systemic response inductors, but since they are not universal, more research processes are needed. To date, the induction of preventive response in plants is achieved with BCA-M to sensitize the host to respond quickly, strong, lasting, and with the least waste of the energy required when a phytopathogen attack occurs (^{Mauch-Mani et al., 2017}). This is because when plants are growing under field conditions, they are continually exposed to different stimuli of abiotic and biotic type (bacteria, fungi, herbivore, oomycetes, viruses, arthropods, among others), which could activate defenses in the plants and influence the improvement of their response capacity by maintaining basal levels of resistance or a biochemical memory (Mauch-Mani et al., 2017). So, when a plant receives the stimulus, changes occur at the physiological, transcriptional, metabolic, and epigenetic levels, an event which is known as a preventive response induction phase for the plant to develop a faster and stronger response to a potential attack, since the plant was already prepared. This kind of preventive resistance can last and be kept during the plant’s life cycle, and the induction of preventive response can be considered transgenerational (Mauch-Mani et al., 2017). However, to understand the molecular mechanisms that induce a preventive response, two potential mechanisms are suggested: i) epigenetic changes in DNA methylation and modification of histones which may carry stress memories and trigger immune responses, and ii) accumulation of kinase proteins activated by mitogens (MPK) (^{Espinas et al., 2016}).

On the other hand, another approach in the study of BCA-M is the use of secondary metabolites (SM), which is considered as a green solution in agriculture since most of those BCA can secrete their metabolites (^{Köhl et al., 2019}). In this way, the bioactive metabolites (polyketides, non-ribosomal peptides, hybrid metabolites of peptide-peptide, terpenes, siderophores, lytic enzymes, among others) are isolated and purified (Figures 1 and 2). These metabolites are of great interest not only for the scientific community but also for the agro-biotechnology industry because of their biological activities against phytopathogens, such as hormones or carriers of chemical elements during biopesticide and biofertilizer production (^{Woo et al., 2014}). SM are highly diverse in chemical structure, but the biosynthesis pathways are related to the primary metabolism, such as the pathway of shikimato acid, acetyl Co-A, sugar derivatives or amino acids (^{Pott et al., 2019}). Understanding these metabolic pathways is very important for the innovation of products focused on phytopathogens control, either using traditional techniques or those labeled as omics.

A classic holistic approach that has been recently re-valuated with the advent of new technologies as metagenomics is the function of suppressive soils, whose microbial community composition keeps the phytopathogens present from infecting or multiplying in the plant of interest. Those soils can be divided into two categories: i) general suppressive soils, with a high content of microbial mass but low suppression levels; and ii) specific suppressive soils, with a high concentration of one or more microbial species and high levels of plant protection against pathogens (^{Mousa and Raizada, 2016}). ^{Shen et al. (2015}) reported that the composition of the microbial community of the rhizosphere of chill pepper plants inoculated with B. amyloliquefaciens (a beneficial bacterium) induced Fusarium suppression and increased the potentially stimulating bacterial diversity (acidobacteria, firmicutes, Leptosphaeria and Phaeosphaeriopsis). However, there are no recent studies about suppressive soils, especially where BCA-M are continually applied for phytopathogen management, so this kind of studies is a challenge for the agro-biotechnological sector focused on ensuring plant health in a sustainable and integrating way by using highly accurate approaches as the omics sciences.

The development and innovation of the agro-biotechnology sector is decisive for the extensive and successful implementation of the BCA-M in Mexico, yet it represents a great scientific challenge. There are multiple studies about BCA-M used in phytopathogens management, but further integrative and creative studies are still needed to develop or innovate profitable, effective, and ecological alternatives to strengthen the use of high-quality and safe biopesticide bioproducts required for plant health in the world.