Investigation of Plant Growth Promotion Potential of Novel Compost‐Derived Bacilli and Anticipating Their In Vivo Effectiveness Through Mathematical Modelling

ABSTRACT

Concerns about the unintended consequences of using pesticides in agricultural settings have led to a shift in practise towards more environmentally friendly methods such as composting, manuring and using plant growth‐promoting bacteria. While compost is known for its ability to promote plant growth and serve as a soil amendment, little is known about the various types of bacteria that can be found in it and how effective they are in sustainable agriculture. This study was designed to assess the potency of a few common isolates from local compost by quantifying their capacity to fix nutrients, produce growth hormone, mitigate abiotic stress, produce lignocellulose‐degrading enzyme and produce soil‐amending enzyme, all while scaling them using an objective in silico method. In addition, the isolates were tested in the field against test crops selected from the monocot and dicot families, namely maize ( Zea mays ) and okra ( Abelmoschus esculentus ). When comparing the treated settings to the untreated control setups, an increase was seen in the quantity of fruits (okra) and percentage of grain filling (maize). By creating a score system, an attempt was made to compare the efficiency of five strains based on the in vitro and in vivo experiments.

Compost derived Bacillus isolates were screened for plant growth promoting traits, scored through modelling and validated in field trials on maize and okra. Selected strains enhanced growth significantly, enabling reduced chemical fertiliser use, ecofriendly bioinputs and improved soil health.

1. Introduction

The rapid intensification of agriculture in recent decades has relied heavily on synthetic fertilisers and chemical pesticides, resulting in ongoing soil degradation, alterations to indigenous microbial ecosystems and heightened environmental and health issues (Lal 2015; Adedeji et al. 2020). Sustainable agriculture practises increasingly prioritise organic amendments and microbial bioinputs to improve soil fertility and diminish reliance on chemicals. Among these, compost application and application of plant growth‐promoting bacteria (PGPB) are two complementary methods that have shown advantages for crop yield and soil health (Bernal et al. 1998). Compost is an organic supplement that has been biologically stabilised by the controlled aerobic breakdown of organic waste, yielding a humified substance that enhances soil structure, nutrient availability and microbial activity (Bernal et al. 1998). The agronomic efficacy of compost is highly variable and contingent on its feedstock content, level of stability and accompanying microbial communities (Bezanson 2014; Nehela et al. 2021). In several emerging agricultural systems, farmers depend on commercially accessible, non‐certified compost obtained from open marketplaces. Local compost materials are often generated under inconsistent, inadequately defined conditions, resulting in considerable variation in nutrient content and microbiological quality. Although physicochemical criteria such as pH and C:N ratio serve as partial markers of compost maturity, they fail to adequately reflect the functional potential or biosafety profile of the resident microbial populations (Seneviratne et al. 2011; Aguilar‐Paredes et al. 2023). High‐throughput sequencing investigations have enhanced our understanding of compost microbiomes at the community level, revealing that Firmicutes, Actinobacteria, Proteobacteria and Bacteroidetes are predominant contributors to composting processes (Aguilar‐Paredes et al. 2023). However, community‐level surveys fail to capture strain‐level functional variability and do not directly guide the selection of microbial candidates suitable for agricultural use. From a pragmatic standpoint, farmers and bioinput developers eventually use specific strains or designated consortia, rendering strain‐level identification and prioritisation crucial. In this context, members of the genus Bacillus have garnered significant interest among compost‐associated bacteria owing to their spore‐forming capability, metabolic adaptability and established functions in nutrient mobilisation, phytohormone synthesis, stress alleviation and pathogen suppression (Mohanty et al. 2021). Many Bacillus sp. have been investigated as biofertilisers or biocontrol agents; however, most studies focus on a single or two characteristics—such as phosphate solubilisation or indole‐3‐acetic acid (IAA) synthesis—without integrating them into a comprehensive evaluation of agronomic significance (Kudoyarova et al. 2019). Furthermore, isolates derived from compost are often documented descriptively, with little attempts to evaluate their comparative efficacy or to correlate in vitro functional characteristics with field performance. Biofilm generation has become a significant bacterial characteristic affecting root colonisation, metabolic coordination and stress tolerance in the rhizosphere (Flemming et al. 2016). The potential to form biofilms has been linked to increased expression of many plant growth‐promoting traits in some Bacillus sp. strains; however, this correlation is not universal and is significantly influenced by strain identity and environmental conditions (Maitra et al. 2022). Thus, biofilm production should be viewed as a contributory and associative characteristic rather than a definitive factor in promoting plant growth in field conditions. Therefore, a significant obstacle in PGPB research is the lack of clear, comparative frameworks that facilitate prioritising strains based on the interplay among numerous functional features. Direct comparison of unprocessed experimental results is often unfeasible because of variations in size, units and biological significance of features. Normalisation‐based multi‐criteria methods, including min–max scaling, are extensively used in data science and agricultural modelling to standardise disparate datasets and facilitate comparison ranking (Kern et al. 2020; Ikram et al. 2022). Notwithstanding their usefulness, these methodologies have rarely been used to assess the efficiency of microbial plant growth promotion in a systematic, repeatable manner.

Therefore, this research aims to functionally identify culturable Bacillus sp. isolates derived from commercially available local compost and to comparatively assess their plant growth‐promoting capability utilising an integrated framework. The objectives were to: (i) isolate and identify prevalent compost‐derived Bacillus sp. strains and evaluate their biosafety‐related phenotypes; (ii) quantify a comprehensive array of plant growth‐promoting traits, including nutrient mobilisation, phytohormone production, stress‐responsive enzymes and biocontrol‐related activities; (iii) assess strain performance in limited field trials utilising maize ( Zea mays L.) and okra ( Abelmoschus esculentus L.) as representative monocot and dicot crops; and (iv) apply a weighted, normalisation‐based multi‐parameter scoring framework to rank strains comparatively based on their overall functional profiles.

This study presents a transparent, literature‐informed prioritising method that combines in vitro functional data with representative in vivo performance, rather than suggesting a predictive or mechanistic model. This study aims to enhance rational selection procedures for compost‐derived PGPB by explicitly recognising strain‐specific variability, environmental influences and methodological constraints, thereby contributing to the wider discussion of sustainable, microbe‐assisted agriculture.

2. Materials and Methods

2.1. Sampling and Physicochemical Characterisation of Compost

Compost samples were obtained from a commercial open market in Kolkata, India (22.5726° N, 88.3639° E), representing non‐certified organic amendments commonly used by small‐scale farmers. To minimise batch variability, compost was collected from three independent vendors and pooled prior to analysis. As production protocols were not standardised, the compost was treated as a representative yet heterogeneous material, and this limitation was taken into account during interpretation. Physicochemical characterisation was conducted following standard protocols. The parameters were measured based on Keen‐Raczkowski Box measurement. The sample pH level was measured on a systronic glass electrode‐bearing pH meter (Jackson 1967). Electric conductivity of the compost sample was measured on systronic E.C meter diluting the sample with distilled water at a ratio of 1:2. The TOC or (total organic carbon) content of the sample was measured by following (Jackson 1967). The available macro and micronutrients in the soil were measured following.

2.2. Isolation and Identification of Culturable Bacterial Strains From Compost Sample

The culturable bacteria were obtained by serial dilution methods using compost samples, spreading on nutrient agar (HiMedia, India). These plates were incubated at 37°C for 24–48 h. After that, individual bacteria were purified by repeated streaks for acquiring axenic cultures. Eighteen isolates of bacteria were obtained using this method based on prevalence. These were stored in nutrient agar slants at 4°C for short‐term preservation.

Isolate selection for subsequent analysis was done only on the basis of antibiotic susceptibility profiling. Susceptibility testing was conducted using the Kirby Bauer disc diffusion test on Mueller Hinton agar using 9 different antibiotics targeting different classes and molecular targets viz. Tetracycline (30 mcg), Vancomycin (30 mcg), Meropenem (10 mcg), Polymixin‐B (300 mcg), Cefalexin (30 mcg), Cotrimoxazole (25 mcg), Ciprofloxacin (5 mcg), Chloramphenicol (10 mcg), Rifampicin (5 mcg) (Hi‐Media). Plates were incubated at 37°C for 24 h, and inhibition zones were measured to determine susceptibility. A laboratory strain of Bacillus sp. was included as a reference control. The susceptibility patterns were evaluated based on Clinical and Laboratory Standards Institute (CLSI) recommendations as appropriate. Isolates that were susceptible to a wide range of antibiotics were considered eligible, whereas isolates that were less susceptible or resistant to more than one antibiotic class were not eligible for further assessment (Rath and Padhy 2015).

2.3. Molecular Identification and Phylogenetic Analysis

Genomic DNA was extracted from bacterial isolates through a standard isolation method, followed by molecular identification via amplification of the 16S rRNA gene using universal primers 27F and 1492R. PCR amplification was performed under conventional cycling conditions, with the resultant amplicons confirmed through agarose gel electrophoresis. To enhance sequencing quality and mitigate mixed‐template interference, cloned 16S rRNA gene products were inserted into the pGEM‐T Easy vector and subsequently sequenced using the Sanger method. The resulting sequences underwent analysis against the NCBI nucleotide database via the BLASTn tool to identify phylogenetic relatives based on sequence similarity. Closely aligned sequences facilitated phylogenetic analysis, which involved multiple sequence alignments and the generation of phylogenetic trees utilising the neighbour‐joining technique in MEGA version 7.0, applying the Kimura two‐parameter model. Bootstrap analysis, consisting of 1000 replications, assessed the resilience of the tree topology. Species‐level identification was achieved through sequence similarity of the 16S rRNA gene and phylogenetic clustering. While 16S rRNA gene analysis provides taxonomic insight, it does not reflect functional gene content or pathogenic potential. The sequences were submitted to the NCBI GenBank database, with accession numbers assigned for each isolate, and phylogenetic characterisation for five isolates was conducted using 16S rRNA gene sequences amplified with primers 27F and 1429R, supplemented by BLAST analysis to match the top 10 identity scores for identification. MEGA 7.0 was employed for the phylogenetic tree generation (Kimura 1980; Kumar et al. 2015; Felsenstein 1985).

2.4. Biofilm Formation and EPS Extraction

2.4.1. Qualitative Biofilm Assessment

Presence of biofilm was detected using Brain Heart Infusion (BHI) Congo Red Agar media. The preparation consisted of BHI broth supplemented with 5% sucrose and 1.5% agar. Congo red dye, at a concentration of 10 g/L, was autoclaved separately under conditions of 15 lb. pressure and 121.6°C for 15 min. After cooling to 55°C, the Congo red indicator was combined with the BHI sucrose solution. Positive results were indicated by the appearance of black, mucoid bacterial colonies after 24–48 h of incubation at 37°C. To ensure accuracy, the experiment was replicated three times (Roy et al. 2021).

2.4.2. Quantitative Biofilm Quantification

For the quantitative assessment of biofilm production, the standard Tissue Culture Plate (TCP) method was employed, utilising two different volumes of 10 and 20 μL. Each of the isolates were cultured in 10 mL of Tryptic Soya Broth (TSB) supplemented with 5% glucose, followed by overnight incubation at 37°C. The overnight cultures were subsequently diluted in a sterile 96‐well polystyrene tissue culture plate, bringing each well's final volume to 200 μL. Bacillus subtilis MTCC 441 served as a positive control, while negative control wells included sterile broth without inoculation. After a 24‐h incubation period at 37°C, the contents of each well were removed and the wells were washed four times with 0.2 mL of phosphate‐buffered saline (PBS) at pH 7.2. To ensure biofilm adherence, a 2% sodium acetate fixative was applied, followed by a staining process using 0.1% crystal violet. Excess stain was rinsed off with deionised water, and the plates were air‐dried. The optical density at 570 nm of the biofilms adhering to the wells was measured using a micro‐ELISA plate reader, with the entire experimental setup replicated three times to minimise errors (Roy et al. 2021).

2.4.3. EPS Extraction and Quantification

To analyse the biochemical components of bacterial biofilm, the methodology outlined by Maitra et al. (2022) was employed. Initially, 2 mL of 18‐h‐old bacterial cultures were inoculated in minimal salt media with 5% glucose (pH 7.1 ± 0.1) and incubated in a rotary shaker at 37°C for 72 h. Following this incubation, the cultures were centrifuged at 1680 g for 10 min, and the supernatant was collected. Cold ethanol was then added in a 1:2 v/v ratio to the supernatant, which was incubated at 4°C for 24 h for precipitation. The biofilm was harvested by centrifuging the matrix sediment at 2419 g for 20 min at room temperature, followed by drying the sediment for 24–48 h at 70°C. The resulting dried powder was collected for further characterisation. The purity of the EPS preparations was qualitatively estimated by the lack of cell debris visible after centrifugation, as well as by low levels of eDNA, indicating that cell breakdown was low during the purification procedure. The biochemical analysis could be used to comparatively evaluate the EPS characteristics of the isolates, rather than for any level of complex molecular characterisation. Uronic acids were not estimated separately in this study.

For quantitative estimation of biochemical components, the carbohydrate concentration in the biofilm was measured using the phenol‐sulphuric acid method as per Dubois et al. The protein concentration was determined spectrophotometrically using Bradford's method. Additionally, extracellular DNA concentration was quantified following Burton's method with the aid of DPA reagent, as referenced by Zulfikar Ali et al.

2.5. Qualitative and Quantitative Assessment of Plant Growth Promoting Traits

2.5.1. Nitrogen‐Related Metabolic Activity

For qualitative detection of nitrogen‐fixing ability, the five isolates were streaked on Jensen's media (Hi‐Media) which was modified by the addition of an indicator bromothymol blue and incubated for 3–8 days at 37°C. The subsequent changes in the media colour from greenish blue (initial) to dark blue (intermediary) indicate different stages of nitrogen fixation to nitrification. Changes in the colour of media from dark blue to yellow (final) indicate ammonification completing the nitrogen fixation cycle (Sulistiyani and Meliah 2017).

2.5.2. Phosphate Solubilisation

For qualitative detection of phosphate solubilisation ability, the five isolates were streaked on Pikovskaya agar media (Hi‐Media) and were incubated for 48 h at 37°C (Paul and Sinha 2016).

For quantitative estimation of the phosphate solubilising abilities of the selected five strains, the spectrophotometric vanadomolybdophosphoric yellow colour standard method was used (Pandea et al. 2019).

2.5.3. Potassium Solubilisation

For qualitative detection of potassium solubilisation ability, the five isolates were streaked on Aleksandrov agar media (Hi‐Media) which was modified by the addition of an acid–base indicator bromothymol blue and were incubated for 3 days at 37°C. As the process of potassium solubilisation takes place, several acids are produced lowering the pH of the media and changing the colour of the media to yellow (Rajawat et al. 2016).

2.6. Phytohormone Production

2.6.1. Indole‐3‐Acetic Acid (IAA)

For quantitative estimation of Indole Acetic acid produced by the selected strains, the isolates were grown in two sets, one set supplemented with 0.1% (w/v) l‐tryptophan in Luria Bertani broth at 28°C for 48 h at a shaking speed of 100 rpm. Salkowski reagent was prepared by mixing 2 mL of 0.5 M FeCl₃ with 49 mL 70% Perchloric acid and 49 mL of water and was stored in an amber bottle (Light sensitive reagent). Freshly prepared Salkowski reagent of 2 mL was added to bacterial supernatant obtained by spinning the 48 h old cultures at 10,000 rpm for 10 min. Samples were incubated till the appearance of pink colour, which was measured at an absorbance of 540 nm. The concentration was measured based on a standard curve prepared in accordance with Sarker and Rashid (2013).

2.6.2. Gibberellic Acid (GA3)

Colorimetric estimation of bacterial Gibberellic acid (GA3) was conducted using a modified Hollbrook method. Bacterial cultures were incubated for 5 and 7 days, followed by mixing 10 mL of absolute alcohol with 15 mL of bacterial cell supernatant and diluting to 40 mL with distilled water. Zinc acetate reagent (prepared with 21.9 g zinc acetate, 1 mL glacial acetic acid and distilled water) was added. After 2 min, potassium ferrocyanide was included, and the total volume was adjusted to 50 mL. After filtering the solution, a 10 mL aliquot was mixed with 8 mL of absolute alcohol and 30% HCl, incubated at 20°C for 75 min and absorbance was measured at 254 nm, with a blank using 5 mL of 5% HCl (Sharma et al. 2021).

2.7. Qualitative and Quantitative Estimation of Stress‐Responsive Enzyme: ACC Deaminase

The selected bacterial isolates were streaked on minimal Dworkin‐Foster (DF) media plates amended with 3 mM ACC as per Kumar et al. (2012). Minimal DF media amended with a regular nitrogen source, that is, ammonium sulphate as a positive control and minimal DF media devoid of any amendment (without nitrogen source) as a negative control was prepared. All plates were incubated for 72 h at 28°C + 1. The presence of ACC deaminase enzyme was further quantified colorimetrically by ninhydrin‐ACC assay by Li et al. (2011). The optical density in the intensity of colour variation after the development of Ruhemann's Purple was measured at 570 nm.

2.8. Enzymatic Activities Related to Soil Health and Biocontrol

A rapid spectrophotometric method was used following the method described by Okyay and Rodrigues (2013) to estimate urease activity. The quantitative estimation of cellulase production was estimated by following Miller (1959) and Sethi et al. (2013). For quantitative estimation of laccase activity, the method described by Sheikhi et al. (2012) was followed. The quantification of enzymatic properties of lignin peroxidase to demethylate methylene blue was conducted spectrophotometrically as per the method described by Jadhav et al. (2016). To estimate amylase activity, the amount of reducing sugar produced by hydrolysis of polysaccharide substrate (starch) by the DNS acid method was estimated (Patel et al. 2019).

Production of catalase enzyme was estimated by a standard method (Iwase et al. 2013). For quantitative estimation of peroxidase activity Kalyani et al. (2011) was followed. The enzymatic activity of protease was measured by using casein as the substrate with crude enzyme extract from selected isolates based on Tsuchida et al. (1986) and Sony and Potty (2016). The enzymatic activity of pectinase was measured as per Shreshtha et al. using DNS substrate. For quantitative estimation of β‐1,3 glucanase activity (Rais et al. 2017) was followed.

2.9. Quantitative Estimation and Classification of Siderophore

The amount of percent siderophore production was quantitatively measured by following (Arora and Verma 2017). The calculation of the percent siderophore unit or psu measurement was done by the formula mentioned below:

$$\text{Percent}\text{siderophore}\text{production}\left(\mathrm{psu}\right)=\frac{\left(A_r-A_s\right)\times 100}{A_r}$$

A _r  = O.D. of CAS (Chrome Azurol S) solution and uninoculated broth at 630 nm. A _s  = O.D. of the sample, CAS solution and cell‐free supernatant.

The selected strains were tested to understand the type of siderophore they are producing. The hydroxamate type of siderophore was detected by using Tetrazolium salt while the catecholate nature of siderophore was detected by following the standard Arnow's Method. Both the experiments were followed by Radhakrishnan et al. (2014).

2.10. Field Trial Design and Plant Growth Assessment

Field experiments were executed using maize ( Zea mays L.) and okra ( Abelmoschus esculentus L.) as exemplars of monocot and dicot crops, respectively. The experimental design was modified from previously established rhizobacterial field application procedures (Roy et al. 2023). Each treatment included three plants per duplicate, with three distinct biological replicates per treatment (n = 9 plants). Treatments were organised according to a randomised block design to reduce spatial variability, with a plant‐to‐plant distance of 1 ft. The crop‐to‐crop distance was maintained at 2 ft. Bacterial inocula were generated by culturing the chosen isolates in nutritional broth for 24 h, thereafter adjusting the cell density to about 10⁶ CFU mL⁻¹. 7 mL (CFU 10⁶/mL) of bacterial culture was centrifuged at 10,000 rpm (rotations per minute) for 15 min. The inocula were administered as aqueous suspensions directly to the rhizospheric area of each plant 15 days post‐sowing. The pellets were half‐diluted with distilled water. Fifteen millilitres of culture was applied in the periphery of the rhizospheric region of each plant. Two subsequent treatments were conducted at 20‐day intervals, ensuring that the final treatment aligned with the pre‐flowering period, as previously detailed (Roy et al. 2023). Post application of treatment, vegetative and reproductive growth parameters were documented and compared with the uninoculated control group.

Quantitative data are presented as mean ± standard deviation (SD). Treatment effects were assessed using one‐way analysis of variance (ANOVA). If significant differences were identified, Tukey's honestly significant difference (HSD) test was used for pairwise comparisons of treatment means. Effect sizes were calculated using eta squared (η ²) to quantify the magnitude of treatment effects. Statistical significance was defined as p < 0.05. All statistical analyses were conducted using standard statistical software in accordance with established analytical procedures (Montgomery 2017).

2.11. Multi‐Parameter Normalisation‐Based Scoring Framework

In order to allow a comparative evaluation of the plant growth‐promoting (PGP) abilities of the isolated bacterial strains, a scoring system based on multiple parameters and normalisation has been established. This method was developed as a heuristic and transparent prioritisation tool in which various experimentally determined characteristics should be united in a single comparative unit, but which would not be used as a predictive model.

2.11.1. Selection of Parameters

Fifteen parameters were quantified in vitro, which were used for scoring. These parameters were selected as they represent different important mechanisms that contribute towards PGP activity, such as biofilm, nutrient mobilisation, phytohormone, stress reduction, siderophore iron acquisition and enzyme activities for soil health, and biocontrol. It was ensured that only parameters measured similarly across all isolates were selected. The importance of multiple traits having a role towards PGP activity is well established, as it is known that collectively, multiple mechanisms have a contribution towards the overall activity, as opposed to a single trait, as explained in Backer et al. (2018) and Ahemad and Kibret (2014).

2.11.2. Data Normalisation

Since the parameters selected varied in scale, units and value range, the raw data were scaled via the min‐max normalisation technique, which was common in the biological and agricultural fields for the treatment of parameters with diverse units and value ranges (Kern et al. 2020; Weraikat et al. 2024). For each parameter, the scaled values were computed using the following equation:

$$\text{Normalized}\text{value}=\frac{x-x_{\min }}{x_{\max }-x_{\min }}$$

where x is the value for a particular parameter over a given isolate, and x _min and x _max are the minimum and maximum values of that parameter, respectively, over all the isolates. This scaling transformed all parameters to have an identical range between 0 and 1, thereby avoiding dominance of the composite score by those parameters with larger absolute values.

2.11.3. Weightage Assignment

Normalised values were weighted according to the relative importance to agriculture that can be inferred in the literature. Greater weight was placed on the ability to form biofilms due to its importance in rhizosphere settlement and the coordination of PGP traits already established in the literature (Ramey et al. 2004; Flemming et al. 2016; Maitra et al. 2022). Nutrient solubilisation properties were weighted considerably for the direct provision of nutrients to plants already established in the literature (Richardson et al. 2009; Adedeji et al. 2020). Phytohormone production, stress enzyme production (ACC deaminase activity) and siderophore production were weighted modestly for the established impacts of these traits on plant growth in a manner that is dependent on the context (Kudoyarova et al. 2019; Glick 2014; Miethke and Marahiel 2007). Enzymatic properties for soil conditioning and biocontrol agents were weighted lower but nonzero due to the supportive yet indirect roles they provide in the promotion of plant growth (Compant et al. 2005; Niu and Li 2022). Although the weighting scheme was literature‐guided and biologically motivated, it was not optimised based on output performance trends to avoid outcome‐driven bias. The sum of all weights was constrained to 100 to facilitate interpretability and proportional contribution assessment. The sum of all weights was constrained to 100 to facilitate interpretability.

2.11.4. Score Calculation and Ranking

A weighted and normalised sum of all parameter values for each isolate was calculated. This created a composite index of its functional capabilities for the tested set of isolates. A higher index value revealed a stronger overall performance capability for isolates based on the chosen characteristics but did not reflect actual, absolute solution efficacy or guaranteed performance in the field. These values have been used strictly for the purpose of isolates ranking in this work.

2.11.5. Sensitivity and Robustness

Sensitivity analysis has also been used to test the robustness of the scoring framework. Individual factor weights were varied by 10% in the positive and negative directions while summing to a constant weight. There was no change in the relative ranking of the isolates under the testing conditions. Lack of sensitivity to variations in the weights has also been advocated for multi‐criteria decision‐making approaches and biological assessment frameworks (Zavadskas et al. 2012).

2.11.6. Scope and Limitations

The proposed scoring system does not address inter‐character associations, environmental fluctuations and host reactions, and it should in no way be considered for its prognostic potential. The application in the context of agriculture and biological data, although established for normalisation‐based multi‐criteria analyses, would necessitate re‐assessment for different biological systems and datasets (Kern et al. 2020).

3. Results

3.1. Basic Characterisation of Compost

The compost utilised in the current experiment was collected from a local market in Kolkata, India (22.5726° N, 88.3639° E). The compost was taken as a commercially available, non‐certified organic matter. The physico‐chemical properties of the compost are presented in Table 1. Physicochemical properties of compost are summed up in Table 1. The pH value of the compost measured 7.5, which falls between the range measured for mature and agronomically sound composts (pH 6.5–8.0). Total organic carbon and total Kjeldahl nitrogen were also measured at 303.1 and 20.5 g/kg, giving the C:N ratio of about 18:1, which falls between the range measured for stabilised composts (15:1–20:1) and yet high to pose minimal risks during soil application with regard to the potential for N immobilisation by the soil biota. Macronutrient analysis showed that there was a considerable amount of both P₂O₅ (19.3 g kg⁻¹) and K₂O (15.6 g kg⁻¹), indicating the potential nutrient‐supplying capability. The compost similarly contained significant amounts of secondary and micronutrients: calcium, magnesium, sulphur, iron, manganese, zinc, copper and boron, all of which were within the ranges of each required for plant growth. Trace levels of cobalt and aluminium were detected and remained within ranges reported in compost‐derived organic amendments.

TABLE 1.

The table illustrates the major physicochemical characteristics of the compost sample obtained locally.

Sample name: Compost‐sample 1
Soil characters	Value
pH	7.5
Organic matter %	48
C:N	18:1
Total organic carbon (g/kg)	303.1
N (g/kg)	20.5
P2O5 (g/kg)	19.3
K2O (g/kg)	15.6
Cu (g/kg)	0.93
Zn (g/kg)	16
Fe (g/kg)	13.3
Mn (g/kg)	10.5
B (g/kg)	17
Ca (g/kg)	22.3
Mg (g/kg)	7.68
S (g/kg)	7
Na (g/kg)	5.8
Co (mg/kg)	10.5
Al‐trace %	0.071

3.2. Isolation, Antibiotic Susceptibility and Selection of Bacterial Isolates

A total of 18 different colonies were isolated on the basis of their prevalence, out of which five were selected, on the basis of differences in colony characters and relative abundance, for further studies (Table 2). These isolates were named as BT, BM, BS, PSB and KSB respectively.

TABLE 2.

Tabular representation zone of inhibition (mm) of the bacterial isolates showing sensitivity towards 9 different antibiotics along with a control strain of Staphylococcus aureus (ATCC 29213).

Isolates	Zone of inhibition (mm) in presence of different antibiotics
	Tetracycline (30 mcg)	Vancomycin (30 mcg)	Meropenem (10 mcg)	Polymixin‐B (300 mcg)	Cefalexin (30 mcg)	Cotrimoxazole (25 mcg)	Ciprofloxacin (5 mcg)	Chloramphenicol (10 mcg)	Rifampicin (5 mcg)
Control strain	26	21	24	26	21	20	26	23	24
TR00K	30	22	33	11	30	32	33	26	18
TR02K	55	40	51	30	36	32	56	38	34
TR03K	31	26	30	27	34	31	34	27	22
TR04K	30	39	46	27	30	28	27	31	27
BA	40	32	43	21	40	42	43	36	28
LCB	48	38	51	35	49	51	43	42	45
BL	52	42	55	39	53	55	57	46	49
BM	31	20	26	15	21	36	30	29	20
BH	48	40	51	29	48	50	51	44	36
BS	27	17	30	14	28	30	32	21	24
BT	30	22	33	11	30	32	33	26	18
TF1	45	34	40	43	48	37	41	39	22
TF3	57	28	58	29	41	53	51	30	36
TF4	42	31	46	32	26	46	59	48	36
TF6	37	22	42	26	33	42	27	37	27
P2	21	39	39	30	37	35	37	29	41
P4	43	24	47	27	21	30	49	47	31
P5	57	33	32	11	34	38	48	28	37
PSB	19	26	20	17	23	30	23	27	25
KSB	18	22	25	10	16	34	28	32	11
KL1	29	46	42	29	27	46	39	37	47
TH1	32	29	31	17	32	41	40	38	38

Antibiotic susceptibility profiling was conducted using a multifaceted panel of antibiotics representing different functional classes and mechanisms of action. The inhibition zone widths for five isolates were expressed as a percentage of the maximum recorded zone for each antibiotic, revealing susceptibility values ranging from 55% to 95%. Isolates BT and BM displayed notably high relative susceptibility, achieving inhibition zones of ≥ 80% for most antibiotics, indicating a strong sensitivity across various mechanisms. In contrast, isolates BS, PSB and KSB showed moderate susceptibility, generally between 60% and 75%, without reduced sensitivity to multiple antibiotic classes. Several non‐selected isolates demonstrated inhibition zones less than 50%, reflecting diminished susceptibility and leading to their exclusion from further analysis. The reference strain of Bacillus sp. showed intermediate susceptibility values, validating the test's reliability. The percentage‐based analysis confirmed that the selected isolates have broad‐spectrum susceptibility to antibiotics targeting various molecular mechanisms, supporting their further functional characterisation in a biosafety context.

3.3. Molecular Identification and Phylogenetic Analysis

For molecular identification, 16S rRNA gene sequence of the strain BT (with 100% query coverage) has shown an identity of 99.80% with Bacillus luti strain MCCC 1A00359; strain BM (with 100% query coverage) has shown an identity of 99.54% with Bacillus wiedmannii strain ROA193; similarly strains BS, PSB and KSB (with their 100% query coverage) have shown 100% identity with Bacillus paramycoides strain MCCC 1A04098, MCCC 1A04098 and MCCC 1A04098 respectively. The sequences were submitted to the NCBI GenBank Database and their accession numbers were obtained, which are given as follows:

• BT: Bacillus luti strain DBBA_BT1 (NCBI acc: MZ229975).

• BM: Bacillus wiedmannii bv. thuringiensis strain BDBA_BM1 (NCBI acc: MZ229894).

• BS: Bacillus paramycoides strain BDBA_SXCM4 (NCBI acc: MW917244).

• PSB: Bacillus paramycoides strain DBBA_P1 (NCBI acc: MZ227489).

• KSB: Bacillus paramycoides strain DBBA_K1 (NCBI acc: MZ227495).

3.4. Biofilm Formation and EPS Extraction

The presence of luxuriant, dark, mucoid colonies was observed in the case of strain BT, while strains BM and BS showed a little lesser degree of luxuriance, darkness and mucus in their colonies (compared to BT). PSB and KSB showed moderate black mucoid growth indicating moderate levels of biofilm production. A standard laboratory strain of Bacillus sp. was used as a reference (Figure 1).

FIGURE 1.

Image shows formation of black mucoid colonies on BHI‐Congo red agar plates indicating positive biofilm formation. Of uninoculated control and strains BT, BM, BS, PSB and KSB respectively.

Biofilm formation was classified based on optical density (OD₅₇₀) values relative to the optical density cut‐off (ODc), following the criteria described by Hassan et al. The ODc was calculated as:

$$\begin{array}{cc}\mathrm{ODc}=\hfill & \text{mean}\mathrm{OD}\text{of the negative control}\hfill \\ \hfill & +3\times \text{standard deviation}\left(\mathrm{SD}\right)\text{of the negative control}\hfill \end{array}$$

Based on this cut‐off value, biofilm formation was categorised as follows:

• OD ≤ ODc: Non‐ or weak biofilm formation.

• ODc < OD ≤ 2 × ODc: Weak biofilm formation.

• 2 × ODc < OD ≤ 4 × ODc: Moderate biofilm formation.

• OD > 4 × ODc: Strong biofilm formation.

The ODc value used in this study was derived from the biofilm measurements of the reference control strain Bacillus subtilis MTCC 441.

Quantitative biofilm analysis showed the presence of variability among the isolates with regard to their potential for biofilm production (Figure 2). Based on the biofilm production capacity classified using the ODc scale described by Hassan et al., it was found that the isolates had moderate to strong biofilm production capacity depending on the inoculum size. From among the isolates, BT and BM were regular producers of biofilms, as they gave values above 4 × ODc, thus having a strong production capacity, irrespective of the inoculum concentration. The moderate producers were isolates BS, PSB and KSB, which were moderate, especially when using the larger amount of inoculum (20 μL), since they fell between 2 × ODc and 4 × ODc.

FIGURE 2.

Graphical representation of the amount of biofilm produced by the 5 selected isolates in comparison to a common laboratory strain of Bacillus sp. where the optical density of biofilm production observed at 570 nm and compared for evaluation of the extent of biofilm production.

The laboratory strain of the reference laboratory strain ( B. subtilis MTCC 441) showed least biofilm formation and was employed only to measure the threshold ODc. This data reveals differences in the amounts of biofilm produced in the isolates and not similar biofilm forming ability in the strains.

The biochemical composition of extracellular polymeric substances (EPSs) from biofilms of five selected isolates was quantitatively analysed, focusing on carbohydrate, protein and extracellular DNA (eDNA) contents (Figure 3). Carbohydrates were the dominant component in EPS, followed by proteins, with eDNA constituting a minor fraction. Among the isolates, BM had the highest EPS carbohydrate content, whereas protein concentrations varied less, with BT and BS having slightly higher levels than PSB and KSB.

FIGURE 3.

The graphical illustration of the concentration (μg/mL) of different components of biofilm of the 5 bacterial strains.

eDNA content was consistently low across all isolates. Despite previous classifications identifying BT and BM as strong biofilm producers, no clear relationship was found between total EPS carbohydrate content and biofilm biomass, indicating that biofilm strength is likely influenced by the organisation and balance of EPS components rather than the concentration of a single component. The study reveals strain‐specific differences in EPS composition, which could impact biofilm architecture and stability. However, the implications of these compositional differences for rhizosphere persistence or plant interactions were not evaluated in this research and remain uncertain.

3.5. Qualitative and Quantitative Assessment of Plant Growth Promoting Traits

3.5.1. Nitrogen‐Related Metabolic Activity

Luxuriant growth was noted in the case of all five strains along with the change in colour of the media due to the production of acid changing the pH of media from bluish green to blue to greenish yellow to yellow indicating the steps of ammonification and nitrification (Figure 4B). BT, BM and BS showed a sharp decrease in pH value after 48 and 72 h of incubation, leading to yellowing of the media indicating rapid nitrification. PSB and KSB showed a slow change in the colour of media indicating ammonification after 48 h and slow nitrification after 72 h of incubation. The Bacillus sp. (reference/control strain) showed no colour change and sparse growth on the nitrogen‐free media, indicating its minimal nitrogen‐fixing abilities.

FIGURE 4.

(A) Potassium solubilisation of 6 strains in modified Aleksandrow‐bromothymol blue media. (B) Nitrogen fixation of the 6 strains in modified Jensen's‐bromothymol blue media.

3.5.2. Phosphate Solubilisation

Luxuriant growth was observed for all five strains along with the control strain on Pikovskyay agar media; however, no distinct clearing zones were observed near the colonies. Therefore, for conclusive determination of the phosphate solubilisation ability of the five strains, the concentration of solubilisation was measured at 430 nm after 72 h of incubation. It was observed that all the bacterial strains, including the control strain, Bacillus sp., have shown comparable concentrations. However, after 120 h of incubation no change in concentration was observed, indicating that all five strains are low to moderate inorganic phosphate solubilisers (Table 3). All experiments were conducted in triplicates to minimise errors.

TABLE 3.

The table illustrates the amount of macro‐nutrients (PO₄), growth‐promoting hormones (IAA and GA₃) and the stress responsive enzyme ACC deaminase produced by the 5 selected strains in comparison to a standard laboratory strain of Bacillus sp.

Bacterial isolates	Macro‐nutrients	Growth promoting hormones				Stress responsive enzymes
	PO4 concentration (μg/mL)	IAA		GA3		ACC deaminase (mmol/L)
		Trp+ (μg/mL)	Trp− (μg/mL)	Concentration at 5 days (μg/mL)	Concentration at 7 days (μg/mL)
BT	15	17.32	14.2	0.21	0.227	0.273
BM	15.3	13.52	12.4	0.230	0.265	0.139
BS	14.25	14.57	10.4	0.227	0.235	0.222
PSB	16.5	9.29	10.6	0.217	0.255	0.192
KSB	14.25	10.35	9.6	0.202	0.234	0.308
Bacillus sp.	10.11	9.8	6	0.123	0.140	0.09

3.5.3. Potassium Solubilisation

The change in colour of the bromothymol blue supplemented Aleksandrow agar media from green to yellow indicated a changing pH of the media due to the production of several inorganic acids including α‐ketogluconic acid, succinic acid, oxalic acid, citric acid, and so forth as the solubilisation of potassium takes place. Among the five bacterial strains, BT and BM showed complete changes in the colour of the media near the halo zone indicating almost complete solubilisation of inorganic potassium (Figure 4A). For the three strains of Bacillus paramycoides BS, PSB and KSB, small halo zones with moderate change in the colour of media were observed. The Bacillus sp. (control strain) showed no colour change and no halo zone on the inorganic potassium media, indicating its non‐potassium solubilisation ability.

3.6. Phytohormone Production

3.6.1. Indole‐3‐Acetic Acid (IAA)

The production of IAA was compared with and without tryptophan (Trp), the precursor of the IAA pathway (Table 3). For BT, BM and BS, higher production was noted in the case of Trp+ conditions (medium amended with tryptophan), whereas moderate production was seen in the case where culture medium does not contain the precursor Tryptophan (Trp−). In the case of KSB, culture medium with or without precursors showed comparable results; the medium containing Trp showed slightly greater value. Only in the case of PSB did the culture medium without precursor show a higher value than the one grown without added precursor. Production of IAA by the control strain, Bacillus sp., was weak when grown without precursor supplementation in the medium and moderate in the presence of supplemented tryptophan. Evidence of tryptophan‐dependent and ‐independent pathways of IAA production both exist in nature. The independent pathway is believed to be associated with the absence of tryptophan in the environment, compelling the bacteria to drive the IAA biosynthetic pathway from some intermediaries of the actual Trp pathway.

3.6.2. Gibberellic Acid (GA3)

GA3 or Gibberellic acid was measured for 5 and 7 days of incubation (Table 3). The sharp trend of increase in the production concentration indicated maximum GA3 production on the 7th day. Production of GA3 on both days was noted to be higher for BM followed by the three strains of Bacillus paramycoides PSB, KSB and BS, and lastly by BT. The production of GA3 by the control strain, Bacillus sp., was found to be moderately less on both days, with a slight increase on the 7th day. All experiments were conducted in triplicates to minimise errors.

3.7. Qualitative and Quantitative Estimation of Stress‐Responsive Enzyme: ACC Deaminase

The growth of DF media by the isolates indicated ACC deaminase activity. No growth was observed in the case of Bacillus sp. laboratory strain. Proper growth was observed by the isolates on normal DF media supplemented with ammonium sulphate as the positive control setup. The strains BT and BM showed proper growth on DF media without any ACC or ammonium sulphate as the nitrogen source, indicating their potency to fix atmospheric nitrogen.

Therefore, to conclusively prove the presence of ACC deaminase, quantification of the activity was tested. ACC deaminase activity was determined colourimetrically by the ninhydrin‐ACC assay. The highest activity was noted for the strain KSB, followed by BT, BS, PSB and BM. The standard Bacillus sp. strain showed the lowest activity at around 0.09 mmol/L indicating the minute presence of this stress‐responsive enzyme. The experiment was conducted thrice to minimise error (Table 3).

3.8. Enzymatic Activities Related to Soil Health and Biocontrol

The urease enzyme production data were recorded at an interval of 3, 5 and 7 days (Table 4). At 0 day of inoculation, the broth was observed to be yellow in colour indicating no urease activity corresponding to the maximum value of absorbance at 430 nm. With increasing hydrolysis of urea, a change in pH along with the change in colour of media was noted. The absorbance at 430 nm was seen to reduce while the peak of 560 nm was seen to increase. For all the strains except for KSB, a sharp declination in the absorbance at 430 nm and an increase in OD₅₆₀ were observed from Day 5. In the case of KSB, the sharp change in media colour and corresponding absorbance value was noted on Day 7.

TABLE 4.

Table illustrates production of various agriculturally important enzymes including enzymes degrading lignocellulolytes, enzymes assisting in biocontrol and siderophores (iron chelating compound) by the 5 selected strains in comparison to a standard laboratory strain of Bacillus sp.

Bacterial isolates	Enzymes for lignocellulosic degradation
	Urease activity						Cellulase (mg/mL)	Laccase (units/mL)	Lignin peroxidase (units of enzyme)	Amylase (units of enzyme)
	Day 3		Day 5		Day 7
	430 nm	560 nm	430 nm	560 nm	430 nm	560 nm
BT	1.349	0.472	1.119	0.592	0.09	3.718	4.0	0.12	0.187	1675
BM	1.467	0.559	1.303	1.292	0.242	3.412	3.3	0.191	0.163	5608
BS	1.327	0.52	1.212	0.701	0.446	0.792	1.6	0.286	0.160	3583
PSB	0.814	0.424	0.65	0.784	0.223	2.448	4.23	0.215	0.243	7491.5
KSB	1.611	0.65	0.409	1.292	0.209	3.452	0.06	0.193	0.188	4666.5
Bacillus sp.	1.012	0.413	0.871	0.621	0.434	0.931	0.03	0.02	0.157	1045

Bacterial isolates	Biocontrol enzymes										Siderophore production
	Catalase (cm)	Pectinase (units of enzyme)	Protease (units of enzyme)	β‐1,3‐Glucanase (mmol/L)	Peroxidase conc						% unit (24 h)	% unit (48 h)
					T1 (mmol/mL)	T2 (mmol/mL)	T3 (mmol/mL)	T4 (mmol/mL)	T5 (mmol/mL)	T6 (mmol/mL)
BT	0.8	1288	734	6.13	0.208	2.604	3.528	4.484	5.044	4.628	53.03	58.12
BM	0.7	1918.3	827	2.53	0.200	1.8	1.988	2.644	2.720	3.436	42.41	40.4
BS	2	506.66	88	6.93	0.204	3.448	3.86	4.488	5.168	5.104	37.04	59.59
PSB	0.8	410	951	5.56	0.204	0.196	2.432	3.124	4.304	4.144	46.300	63.75
KSB	0.8	605	89	3.32	0.200	0.216	2.532	3.972	4.932	5.004	55.430	31.54
Bacillus sp.	0.6	359	647	2.14	0.200	0.223	0.452	1.213	1.672	2.014	21.9	30.08

In the case of other lignocellulolytic enzyme clusters (Table 4) the three strains of Bacillus paramycoides, mainly PSB (Bacillus paramycoides strain DBBA_P1), showed the highest activity in cellulase, lignin peroxidase and amylase. For laccase activity, the Bacillus paramycoides strain BS (strain BDBA_SXCM4) showed the maximum activity. The two strains BT and BM showed moderately comparable results in all 4 cases. The detailed results are described in Table 4. All the enzymes were estimated in comparison to a standard Bacillus sp. All experiments were conducted in triplicates to minimise error.

The results obtained with different biocontrolling enzyme clusters were shown in Table 4. As per the activities of the biocontrol enzyme cluster, the strain Bacillus paramycoides BS showed better results in the case of catalase, beta‐1,3‐glucanase and peroxidase. The strain PSB showed the highest activity in the case of protease activity and the second highest beta‐1,3‐glucanase activity. The other three strains showed comparable results. The highest pectinase activity was noted in the case of BM followed by BT; the three strains of Bacillus paramycoides showed comparable activity. The control Bacillus sp. strain was used for comparison for all the enzymatic assays. All the experiments were conducted in triplicates to minimise error.

3.9. Quantitative Estimation and Classification of Siderophore

Siderophores are iron‐chelating compounds acting as one of the most important metabolites for both plant growth promotion and control against phytopathogens. The iron chelating compounds are measured in terms of percent siderophore unit or psu at 24 and 48 h of incubation (Table 4). Based on % siderophore unit, BS, PSB and BT showed higher value of psu after 48 h of incubation, while for the other three bacterial isolates 24 h of incubation showed greater value. The iron‐chelating compounds of the standard laboratory species of Bacillus sp. were found to be low in comparison to the five test isolates. The experiment was conducted thrice to minimise the chances of errors. The nature of the siderophore produced was analysed to detect the type of siderophore produced. All five test strains showed orange to pink media coloration indicating the presence of a hydroxamate type of siderophore. None of the strains showed catechol type or mixed type of siderophore formation.

3.10. Field Trial Design and Plant Growth Assessment

To evaluate the in vivo effects of the selected bacterial isolates, a pilot‐scale field experiment was conducted using maize ( Zea mays L.) and okra ( Abelmoschus esculentus L.). Plant growth responses were assessed by recording two vegetative parameters (plant height and number of leaves) and two reproductive parameters for each crop following completion of three inoculation treatments. Data from inoculated treatments were compared with an uninoculated control (Figure 5A,B).

FIGURE 5.

(A) Physiological parameters of Zea mays L. recorded showing the 5 bacterial strains were used as treatment setup in comparison to the untreated control setup. (B) Physiological parameters of Abelmoschus esculentus L. recorded showing 5 bacterial strains were used as treatment setup in comparison to the untreated control setup.

One‐way ANOVA revealed significant treatment effects on all measured vegetative and reproductive parameters of maize, including plant height (F _5,12 = 374.18, p < 0.001), leaf number (F _5,12 = 35.73, p < 0.001), cob length (F _5,12 = 39.61, p < 0.001), cob weight (F _5,12 = 583.61, p < 0.001) and grain filling percentage (F _5,12 = 566.82, p < 0.001) (Table 5). Tukey's HSD post hoc analysis indicated that BT‐ and BM‐treated plants exhibited the highest vegetative growth and reproductive performance across most parameters, followed by PSB and BS, whereas KSB and the uninoculated control showed comparatively lower responses.

TABLE 5.

Tabular representation of one‐way ANOVA showed significant treatment effects for all parameters (p < 0.001).

Treatment	Plant height (in.)	Leaves (no.)	Cob length (cm)	Cob weight (g)	Grain filling (%)
BT	72.17 ± 0.76 a	14.33 ± 1.15 a	23.00 ± 1.20 a	318.03 ± 11.30 a	91.30 ± 0.89 a
BM	67.67 ± 3.06 b	13.67 ± 1.15 a	21.53 ± 0.50 b	288.70 ± 6.83 b	82.67 ± 1.53 b
PSB	42.63 ± 1.05 c	11.33 ± 0.58 b	18.33 ± 1.15 c	156.23 ± 6.37 c	58.80 ± 2.44 c
BS	39.67 ± 2.08 d	9.33 ± 0.58 c	18.17 ± 1.76 c	144.10 ± 7.69 d	59.93 ± 0.50 c
KSB	26.93 ± 1.40 e	8.67 ± 0.58 c	17.50 ± 0.50 d	111.17 ± 3.75 e	42.70 ± 0.82 d
Control	25.50 ± 1.17 e	7.33 ± 0.58 d	12.13 ± 0.30 e	66.67 ± 4.51 f	41.50 ± 1.80 d

Note: Tukey's HSD test was applied for post hoc comparisons for Zea mays for the 5 treatment setups and uninoculated control. The presence of superscript letters such as a, b, c, d, and e in scientific tables represents a standardized statistical convention known as a Compact Letter Display (CLD). This notation serves as a visual shorthand for the results of multiple pairwise comparison tests often referred to as post‐hoc tests that follow an initial Analysis of Variance (ANOVA). The primary function of these letters is to identify which specific treatment means differ significantly from one another while controlling for the inflation of Type I error rates that naturally occurs when making multiple simultaneous comparisons.

One‐way ANOVA revealed significant treatment effects on all vegetative and reproductive parameters of okra, including plant height (F _5,12 = 179.91, p < 0.001), leaf number (F _5,12 = 495.41, p < 0.001), fruit length (F _5,12 = 161.54, p < 0.001) and fruit weight (F _5,12 = 516.19, p < 0.001) (Table 6). Tukey's HSD post hoc analysis indicated that BT‐treated plants exhibited the highest values across all growth parameters, followed by BM. PSB and BS produced intermediate responses, whereas KSB and the uninoculated control showed comparatively lower growth and yield‐related traits.

TABLE 6.

Tabular representation of one‐way ANOVA showed significant treatment effects for all parameters (p < 0.001).

Treatment	Plant height (in.)	Leaves (no.)	Fruit number	Fruit length (cm)	Fruit weight (g)
BT	48.07 ± 0.25 a	14.70 ± 0.10 a	47.67 ± 0.58 b	21.80 ± 0.10 a	26.00 ± 1.00 a
BM	46.10 ± 0.30 b	13.30 ± 0.10 b	49.50 ± 0.10 a	19.00 ± 1.00 b	22.80 ± 0.10 b
PSB	42.57 ± 0.40 c	12.13 ± 0.25 c	43.70 ± 0.10 c	18.13 ± 0.15 c	19.50 ± 0.10 c
BS	39.43 ± 0.21 d	8.20 ± 0.17 e	36.20 ± 0.53 d	11.50 ± 0.10 e	12.80 ± 0.10 e
KSB	35.77 ± 0.15 e	9.10 ± 0.10 d	34.60 ± 0.10 e	15.00 ± 1.00 d	18.20 ± 0.10 d
Control	35.67 ± 1.53 e	7.60 ± 0.44 f	17.33 ± 0.58 f	11.07 ± 0.15 e	12.00 ± 0.10 e

Note: Tukey's HSD test was applied for post hoc comparisons for Abelmoschus esculentus for the 5 treatment setups and uninoculated control. The presence of superscript letters such as a, b, c, d, and e in scientific tables represents a standardized statistical convention known as a Compact Letter Display (CLD). This notation serves as a visual shorthand for the results of multiple pairwise comparison tests often referred to as post‐hoc tests that follow an initial Analysis of Variance (ANOVA). The primary function of these letters is to identify which specific treatment means differ significantly from one another while controlling for the inflation of Type I error rates that naturally occurs when making multiple simultaneous comparisons.

3.11. Multi‐Parameter Normalisation‐Based Scoring Framework

The multi‐parameter normalisation‐based scoring framework demonstrated distinct performance stratification among the five selected bacterial isolates (Figure 6; Table 7). BT achieved the highest composite score (281.86), indicating superior overall functional potential across the evaluated plant growth–promoting traits. BM, BS and PSB followed in descending order, while KSB recorded the lowest composite score (64.61), reflecting comparatively weaker functional performance. High‐ranking isolates, particularly BT and BM, consistently demonstrated elevated contributions across multiple functional categories, including biofilm formation, nutrient mobilisation, phytohormone production, stress‐responsive enzyme activity and agriculturally relevant extracellular enzyme production. In contrast, KSB exhibited limited contributions across most parameters, resulting in substantially lower cumulative scores. Biofilm formation represented the most influential trait category in the scoring framework, receiving the highest individual weight assignments (15.0 each for 10 and 20 μL assays). This weighting was evident in the graphical pie‐chart representations, where BT and BM displayed dominant proportional contributions from biofilm‐associated features, underscoring their strong rhizosphere colonisation potential. Secondary contributions originated from nutrient solubilisation traits (phosphate mobilisation), phytohormone production (IAA and gibberellic acid), siderophore‐mediated iron acquisition and stress‐associated enzymatic activities. Enzymatic clusters associated with soil conditioning and biocontrol potential, including cellulase, protease, amylase, pectinase, glucanase, urease, catalase and peroxidase activities, contributed additively to the composite scores. Although these traits received lower individual weights, their cumulative influence further distinguished moderate‐ and low‐performing isolates. Overall, the scoring framework facilitated integrated ranking of isolates by simultaneously accounting for multiple functional attributes. The consistent dominance of BT across the majority of weighted traits supports its prioritisation as the most promising candidate for subsequent field evaluation, followed by BM and BS. In contrast, PSB and KSB demonstrated comparatively limited composite functional performance.

FIGURE 6.

Graphical representation of scores obtained by each of the 5 bacterial isolates based on a machine‐generated unique scoring system.

TABLE 7.

Table illustrates the parameters taken, weights given, feature numbers allotted and scores obtained for this min‐max scoring system.

Parameters	Corresponding features	Allotted weightage	Scores obtained by the bacterial strains as generated by scoring system
			BT	BM	BS	PSB	KSB
Biofilm10microlit	Feature0	15.0	75.0	46.32	37.50	0.0	0.0
Biofilm20microlit	Feature1	15.0	75.0	45.65	50.54	16.30	0.0
Cellulase	Feature2	2.0	9.54	7.76	3.69	10.0	0.0
Protease	Feature3	2.0	7.48	8.56	0.0	10.0	0.01
Pectinase	Feature4	2.0	5.28	10.0	0.64	0.0	1.29
Amylase	Feature5	2.0	0.0	6.76	3.28	10.0	5.14
Beta1,3glucanase	Feature6	2.0	8.18	0.0	10.0	6.88	1.79
Catalase	Feature7	2.0	0.76	0.0	10.0	0.76	0.76
ACCdeaminase	Feature8	5.0	19.82	0.0	12.27	7.84	25.0
Laccase	Feature9	2.0	0.0	6.42	10.0	7.33	6.50
IAATrp+	Feature10	3.0	15.0	7.90	9.86	0.0	1.98
IAATrp−	Feature11	3.0	15.0	9.13	2.60	3.26	0.0
GA3at5days	Feature12	2.5	3.57	12.5	11.16	6.69	0.0
GA3at7days	Feature13	2.5	0.0	12.5	2.63	9.21	2.30
PO4	Feature14	10.0	16.66	23.33	0.0	50.0	0.0
% siderophoreunits (24 h)	Feature15	2.0	8.69	2.92	0.0	5.03	10.0
% siderophoreunits (48 h)	Feature16	2.0	8.25	2.75	8.70	10.0	0.0
Urease (430 nm)‐Day3	Feature17	0.33	1.11	1.365	1.07	0.0	1.66
Urease (430 nm)‐Day5	Feature18	0.33	1.38	1.66	1.52	0.021	0.0
Urease (430 nm)‐Day7	Feature19	0.33	0.17	0.81	1.66	0.067	0.0
Urease (560 nm)‐Day3	Feature20	0.33	0.0	0.09	0.051	1.66	0.191
Urease (560 nm)‐Day5	Feature21	0.33	0.0	0.43	0.067	1.35	1.666
Urease (560 nm)‐Day7	Feature22	0.33	1.66	1.49	0.0	1.51	1.51
PeroxidaseConcT1	Feature23	0.33	1.66	0.0	0.833	0.83	0.0
PeroxidaseConcT2	Feature24	0.33	1.23	0.822	1.66	0.0	0.01
PeroxidaseConcT3	Feature25	0.33	1.37	0.0	1.66	0.39	0.48
PeroxidaseConcT4	Feature26	0.33	1.66	0.0	1.66	0.43	1.20
PeroxidaseConcT5	Feature27	0.33	1.58	0.0	1.66	1.07	1.50
PeroxidaseConcT6	Feature28	0.33	1.19	0.0	1.66	0.70	1.56

4. Discussion

This study assessed compost‐derived Bacillus sp. isolates through an integrated framework that combined in vitro functional screening, statistically supported field trials and a normalisation‐based composite ranking approach. Although previous research has documented individual plant growth–promoting (PGP) traits in compost‐associated bacteria, few studies have systematically connected multi‐trait functional profiling with field performance outcomes. Consequently, this work offers a comparative, performance‐oriented prioritisation strategy for candidate strains.

Physicochemical characterisation of the compost revealed a near‐neutral pH (7.5) and a C:N ratio of approximately 18:1, values typically associated with stabilised organic amendments that support microbial activity and nutrient mineralisation (Bernal et al. 1998; BrustGerald 2019). While compost maturity was not evaluated using respirometric or humification indices, the measured parameters are within agronomically suitable ranges. In this study, compost was primarily utilised as a microbial reservoir rather than as a fertiliser input, highlighting its role as a source of functionally diverse bacteria.

The predominance of Firmicutes, particularly Bacillus species, among the isolated strains aligns with compost microbiome surveys that report selective enrichment of spore‐forming taxa under thermophilic, aerobic composting conditions. Members of the Bacillus cereus sensu lato group, such as B. paramycoides, B. luti and B. weidmannii, remain relatively underexplored in agricultural contexts despite their metabolic versatility and environmental resilience. Their isolation in this study reinforces the emerging view of compost‐derived Bacilli as important contributors to soil functional diversity. However, the phylogenetic proximity of certain isolates to opportunistic pathogens necessitates careful evaluation. Antibiotic susceptibility profiling was used as an initial biosafety screening step to exclude strains with reduced susceptibility across multiple antibiotic classes. This method does not replace comprehensive genomic screening. Future applications of these strains should confirm the absence of toxin‐associated gene clusters and assess horizontal gene transfer risks through whole‐genome sequencing and comparative genomics (Ehling‐Schulz et al. 2015).

Functional trait profiling showed that all selected isolates exhibited multiple PGP‐related mechanisms, including nutrient solubilisation, phytohormone synthesis, siderophore production, stress‐responsive enzyme activity and extracellular enzymatic functions related to organic matter turnover. This multifunctionality is consistent with current understanding that effective plant growth promotion arises from synergistic interactions among complementary microbial traits rather than from isolated mechanisms (Backer et al. 2018; Ahemad and Kibret 2014).

Biofilm formation was a prominent functional attribute among high‐performing isolates. Biofilms enhance rhizosphere colonisation, increase microbial persistence and facilitate metabolic coordination within microbial communities (Ramey et al. 2004; Flemming et al. 2016). The relationship between biofilm formation and plant performance is, however, context‐dependent. Factors such as soil heterogeneity, plant root exudate composition, moisture availability and microbial competition can influence biofilm stability and functional expression in situ. While strong biofilm‐forming isolates performed well in this study, biofilm production alone should not be considered a universal predictor of field efficacy. Field validation with maize and okra demonstrated statistically significant improvements in both vegetative and reproductive traits following bacterial inoculation. High‐performing strains (BT and BM) consistently enhanced yield‐associated parameters, including grain‐filling percentage and fruit number, as well as vegetative growth indicators. Although the magnitude of response differed between crops, reflecting physiological and developmental differences between monocots and dicots, treatments were analysed independently to prevent overgeneralisation. The consistent ranking of top‐performing strains across both crops indicates functional robustness across plant systems rather than narrow host specificity. The normalisation‐based scoring framework offered a structured method for integrating heterogeneous experimental datasets into a single comparative index. In contrast to machine learning‐based prediction models, which require large training datasets and often lack interpretability, this framework prioritised transparency and biological relevance. Trait weighting was guided by published agronomic evidence rather than optimised for outcome performance, reducing model bias. Sensitivity analysis demonstrated that isolate rankings remained stable under moderate weight perturbations, supporting the robustness of the approach (Zavadskas et al. 2012).

Despite these advantages, the scoring framework has inherent limitations. It does not account for nonlinear interactions among traits, environmental modulation of microbial performance, or plant genotype‐specific responses. Furthermore, the framework is most appropriate for comparative prioritisation rather than predictive modelling. Future research could combine this approach with multivariate statistical models or machine learning classifiers to enhance predictive resolution while maintaining interpretability.

From an applied perspective, compost‐derived Bacilli sp. present several advantages for microbial inoculant development, including spore‐based formulation stability and broad metabolic capacity. Integrating microbial inoculants with organic amendments, such as compost, may enhance soil fertility by providing both nutrient inputs and biological stimulation. However, effective field deployment requires optimising inoculum formulation, delivery methods, carrier compatibility and long‐term persistence across diverse agronomic conditions. Ecological considerations are essential. Introducing selected microbial strains into agricultural soils may alternative microbial community structure and function. Long‐term monitoring of soil microbiome dynamics and ecosystem impacts is necessary to ensure ecological compatibility and sustainability (Compant et al. 2005). Incorporating metagenomic analysis and multi‐season field trials will further enhance translational relevance.

In summary, this study demonstrates that compost ecosystems are valuable reservoirs of agriculturally relevant microbial diversity. By integrating functional screening, field validation and normalisation‐based ranking, this work establishes a systematic framework for the rational selection of plant growth–promoting bacteria. Although further genomic validation and ecological assessment are necessary before large‐scale application, the methodological strategy presented here advances data‐driven, reproducible microbial inoculant development for sustainable agriculture.

5. Conclusion

This study demonstrates that compost ecosystems are valuable deposits of multifunctional Bacillus strains with quantifiable plant growth–promoting potential. Through the integration of comprehensive in vitro functional screening and statistically validated field trials in maize and okra, this research establishes a systematic framework for evaluating and prioritising microbial candidates for agricultural application. The observed enhancements in vegetative growth and reproductive performance in both monocot and dicot crops underscore the functional significance of selected compost‐derived isolates under practical cultivation conditions. A key contribution of this study is the development and implementation of a normalisation‐based multi‐parameter scoring framework, which enables transparent, reproducible and biologically interpretable ranking of isolates. This methodology facilitates the simultaneous evaluation of diverse functional traits and reduces bias inherent in single‐parameter screening. Furthermore, the stability of rankings under sensitivity analysis demonstrates the robustness of the composite evaluation strategy. Although the results demonstrate strong functional potential for selected isolates, particularly BT and BM, large‐scale agricultural implementation will require additional validation. Essential next steps include genome‐level biosafety assessment, extended field trials, formulation optimisation and ecological impact analysis. Nonetheless, the integrative methodology presented provides a practical framework for connecting laboratory‐scale microbial screening with field‐level performance evaluation. In summary, this research advances the data‐driven development of microbial bioinoculants and supports the broader goal of promoting sustainable, biologically based agricultural practises through the strategic use of compost‐associated microbial diversity.

Author Contributions

D.M.: conceptualisation, research and preparation of draft. B.R.: conceptualisation, research and review of the original draft. A.C. and S.M.: design of machine learning‐based models. B.D.: conceptualisation, isolation, basic characterisation and in vivo studies. I.C.: conceptualisation and supervision. R.C.: supervision. A.K.M. and A.K.: planning, conceptualisation, supervision and acquisition of fund. S.S.C.: supervision.

Funding

This work was supported by the Department of Biotechnology, Ministry of Science and Technology, India (BT/INF/22/SP41296/2020).

Ethics Statement

The authors have nothing to report.

Conflicts of Interest

The authors declare no conflicts of interest.

Data Availability Statement

All materials are reported in the text, and all data collected are reported in the manuscript.