Search

SynerGNet is a novel approach to predicting drug synergy against cancer cell lines. In this study, we discuss in detail the construction process of SynerGNet, emphasizing its comprehensive design tailored to handle complex data patterns. Additionally, we investigate a counterintuitive phenomenon when integrating more augmented data into the training set results in an increase in testing loss alongside improved predictive accuracy. This sheds light on the nuanced dynamics of model learning. Further, we demonstrate the effectiveness of strong regularization techniques in mitigating overfitting, ensuring the robustness and generalization ability of SynerGNet. Finally, the continuous performance enhancements achieved through the integration of augmented data are highlighted. By gradually increasing the amount of augmented data in the training set, we observe substantial improvements in model performance. For instance, compared to models trained exclusively on the original data, the integration of the augmented data can lead to a 5.5% increase in the balanced accuracy and a 7.8% decrease in the false positive rate. Through rigorous benchmarks and analyses, our study contributes valuable insights into the development and optimization of predictive models in biomedical research.

Abstract Combination therapy has gained popularity in cancer treatment as it enhances the treatment efficacy and overcomes drug resistance. Although machine learning (ML) techniques have become an indispensable tool for discovering new drug combinations, the data on drug combination therapy currently available may be insufficient to build high-precision models. We developed a data augmentation protocol to unbiasedly scale up the existing anti-cancer drug synergy dataset. Using a new drug similarity metric, we augmented the synergy data by substituting a compound in a drug combination instance with another molecule that exhibits highly similar pharmacological effects. Using this protocol, we were able to upscale the AZ-DREAM Challenges dataset from 8798 to 6,016,697 drug combinations. Comprehensive performance evaluations show that ML models trained on the augmented data consistently achieve higher accuracy than those trained solely on the original dataset. Our data augmentation protocol provides a systematic and unbiased approach to generating more diverse and larger-scale drug combination datasets, enabling the development of more precise and effective ML models. The protocol presented in this study could serve as a foundation for future research aimed at discovering novel and effective drug combinations for cancer treatment.

Drug combination therapy shows promise in cancer treatment by addressing drug resistance, reducing toxicity, and enhancing therapeutic efficacy. However, the intricate and dynamic nature of biological systems makes identifying potential synergistic drugs a costly and time-consuming endeavor. To facilitate the development of combination therapy, techniques employing artificial intelligence have emerged as a transformative solution, providing a sophisticated avenue for advancing existing therapeutic approaches. In this study, we developed SynerGNet, a graph neural network model designed to accurately predict the synergistic effect of drug pairs against cancer cell lines. SynerGNet utilizes cancer-specific featured graphs created by integrating heterogeneous biological features into the human protein–protein interaction network, followed by a reduction process to enhance topological diversity. Leveraging synergy data provided by AZ-DREAM Challenges, the model yields a balanced accuracy of 0.68, significantly outperforming traditional machine learning. Encouragingly, augmenting the training data with carefully constructed synthetic instances improved the balanced accuracy of SynerGNet to 0.73. Finally, the results of an independent validation conducted against DrugCombDB demonstrated that it exhibits a strong performance when applied to unseen data. SynerGNet shows a great potential in detecting drug synergy, positioning itself as a valuable tool that could contribute to the advancement of combination therapy for cancer treatment.

Computational modeling is an essential component of modern drug discovery. One of its most important applications is to select promising drug candidates for pharmacologically relevant target proteins. Because of continuing advances in structural biology, putative binding sites for small organic molecules are being discovered in numerous proteins linked to various diseases. These valuable data offer new opportunities to build efficient computational models predicting binding molecules for target sites through the application of data mining and machine learning. In particular, deep neural networks are powerful techniques capable of learning from complex data in order to make informed drug binding predictions. In this communication, we describe Pocket2Drug, a deep graph neural network model to predict binding molecules for a given a ligand binding site. This approach first learns the conditional probability distribution of small molecules from a large dataset of pocket structures with supervised training, followed by the sampling of drug candidates from the trained model. Comprehensive benchmarking simulations show that using Pocket2Drug significantly improves the chances of finding molecules binding to target pockets compared to traditional drug selection procedures. Specifically, known binders are generated for as many as 80.5% of targets present in the testing set consisting of dissimilar data from that used to train the deep graph neural network model. Overall, Pocket2Drug is a promising computational approach to inform the discovery of novel biopharmaceuticals.

Introduction: Gap-free/continuous cement margins have been considered important for the longevity of indirect dental restorations. Bacterial species have demonstrated esterase-like activity that can cause biodegradation of resin composites. Aim: The aim of this study was to evaluate the effect of the esterase-like activity of E. faecalis and L. casei on three resin luting cements. Settings and Design: In-vitro study materials and three resin luting cements tested were: Variolink N, Rely X U200 and Panavia F2.0. E. faecalis and L. casei suspensions and supernatants were assessed for enzymatic activity by bacterial esterase activity assay. Circular samples of resin luting cements were exposed to suspensions of E. faecalis and L. casei for 7 and 28 days followed by testing for solubility, microhardness and bishydroxy propoxy phenyl propane (BisHPPP) release. Results: E. faecalis and L. casei both demonstrated esterase-like activity. Bacterial suspensions had significantly increased enzymatic activity than supernatant solutions (P < 0.05). There was no significant reduction in microhardness or increased weight loss in all three cements after incubation in E. faecalis and L. casei for 7 and 28 days. BisHPPP release signifying resin degradation was seen after 7 and 28 days of incubation in E. faecalis and L. casei. Conclusion: Within the limitations of this in vitro study, E. faecalis and L. casei demonstrated esterase-like activity. BisHPPP release was evident in all three cements after 7 and 28 days. However, the bacterial strains did not significantly reduce the microhardness or cause weight loss of the tested resin luting cements (Variolink N, Panavia F2.0 and Rely X U200) after 7 and 28 days of incubation.

The binding of small organic molecules to protein targets is fundamental to a wide array of cellular functions. It is also routinely exploited to develop new therapeutic strategies against a variety of diseases. On that account, the ability to effectively detect and classify ligand binding sites in proteins is of paramount importance to modern structure-based drug discovery. These complex and non-trivial tasks require sophisticated algorithms from the field of artificial intelligence to achieve a high prediction accuracy. In this communication, we describe GraphSite, a deep learning-based method utilizing a graph representation of local protein structures and a state-of-the-art graph neural network to classify ligand binding sites. Using neural weighted message passing layers to effectively capture the structural, physicochemical, and evolutionary characteristics of binding pockets mitigates model overfitting and improves the classification accuracy. Indeed, comprehensive cross-validation benchmarks against a large dataset of binding pockets belonging to 14 diverse functional classes demonstrate that GraphSite yields the class-weighted F1-score of 81.7%, outperforming other approaches such as molecular docking and binding site matching. Further, it also generalizes well to unseen data with the F1-score of 70.7%, which is the expected performance in real-world applications. We also discuss new directions to improve and extend GraphSite in the future.

Abstract Background There are no previous reports of paliperidone palmitate's (PP) long term tolerability or pharmacokinetics of the highest dose in patients with schizophrenia. This study evaluates safety and tolerability, as well as pharmacokinetics, of the highest marketed dose of PP (150 mg eq. [234 mg]) in stable patients with schizophrenia over a 1-year period. Methods In this 1-year prospective study, eligible patients (aged 18-65 years; Positive and Negative Syndrome Scale's total score ≤ 70) received an initial deltoid injection of PP 150 mg eq. The second injection one week later and subsequent once-monthly injections were deltoid or gluteal. All injections were to be PP 150 mg eq. Patients willing to participate in intensive pharmacokinetic sampling were classified as Treatment A. Patients unwilling to undergo intensive pharmacokinetic sampling or unable to tolerate the 150 mg eq. dose (consequently receiving flexible doses of 50, 100 or 150 mg eq.) were classified as Treatment B. Results Of the 212 patients (safety analysis set), 73% were men; 45% white; 20% black; 34% Asians; mean (SD) age 41 (10.2) years, and mean (SD) baseline Positive and Negative Syndrome Scale total score 54.9 (9.03). A total of 53% (n = 113) patients completed the study and 104 received PP 150 mg eq. throughout. Mean (SD) mode dose of PP was 144.8 (19.58) mg eq. The dosing initiation regimen resulted in rapidly achieved and maintained therapeutic paliperidone levels over the study (average concentrations during the dosing interval were 34.7, 40.0, and 47.8 ng/mL after the 2nd, 8th, and 14th injection respectively). Most frequent (≥ 10%) treatment-emergent adverse events were nasopharyngitis (n = 37), insomnia (n = 32), injection-site pain (n = 32), headache (n = 28), and tachycardia (n = 27). Akathisia (n = 19) and tremor (n = 11) were the most common extrapyramidal adverse events. 33 patients had an SAE and 27 discontinued due to treatment-emergent adverse events. No deaths were reported. Mean (SD) weight change from baseline was 2.5 (5.41) kg at endpoint. Patients' psychoses remained stable. Conclusions Safety results after one-year therapy with the highest available dose of once-monthly paliperidone palmitate were consistent with results from previous studies, with no new concerns noted. Plasma concentrations were within the expected range. Trial registration no ClinicalTrials.gov: NCT01150448

Primary dysmenorrhea denotes the onset of recurrent lower abdominal pain and uterine contractions throughout the bleeding phase of menstruation in the absence of any underlying pelvic pathology. Core stability exercise is considered a beneficial exercise program for managing several health problems. Various relaxation techniques can be used in womanhood with primary dysmenorrhea for relief of pain and improving their quality of life (QOL). The aim of this study was to learn and understand the benefits of core stabilization exercise along with relaxation techniques for primary dysmenorrhea in an unmarried girl. This single-case study was planned to analyze the effect of core stability exercise along with relaxation techniques in the management of primary dysmenorrhea symptoms. Relaxation exercises along with core stability exercises were given to subjectswith primary dysmenorrhea for 20 sessions of 30 minutes spanned for 5 weeks, four sessions a week. Along with the demographic profile, pre- and post -intervention value of pain in the visual analog scale (VAS) and Working Ability, Location, Intensity, Days of Pain, Dysmenorrhea (WaLIDD) score was obtained, recorded, and analyzed. This single-case study results showed significant improvement in the outcome of pain in the VAS and WaLIDD score after the intervention of core stability exercise along with relaxation exercise for the primary dysmenorrhea patients. Pre- and post-intervention of core stability exercise along with Mitchell's relaxation exercises, measurements of VAS, and dysmenorrhea severity in WaLIDD scores revealed an effective reduction in pain and severity using core stability training and relaxation training in an unmarried girl diagnosed with primary dysmenorrhea., Competing Interests: There are no conflicts of interest., (Copyright: © 2024 Journal of Education and Health Promotion.)

The purification and biochemical characterization of the extracellular alpha amylase from A.tamarii MTCC5152 were studied. The combined use of ion exchange and gel filtration chromatographic methods were used for purification studies. The specific activity was significantly increased (33 fold) and 19.41 fold purification of the enzyme α -amylase with 24% yield was achieved. The enzyme had an optimal pH of 6.5 and exhibited its highest activity at 55 °C. It is active over a wide range of pH 5-7 at room temperature. The enzyme is relatively stable in the temperature range of 25-35 °C for a period of 4 h hence, more suitable for industrial applications. K m and V max value of the enzyme was to be 5.882 mg/mL and 0.803 mg/mL/min respectively using starch as the substrate. The purified protein showed a single band on native and SDS PAGE and the molecular weight was found to be 31 kDa. Starch zymogram also revealed one clear zone of amylolytic activity which corresponded to the band obtained with native PAGE and SDS/PAGE. The characterization studies showed that the enzyme activity is inhibited by Ca 2+ , Mn 2+ , Hg 2+ , Fe 2+ .

A hybrid piezo/triboelectric nanogenerator (H/P-TENG) is designed for mechanical energy harvesting using polymer ceramic composite films; polydimethylsiloxane/Ba(Zr 0.2 Ti 0.8 )O 3 -0.5(Ba 0.7 Ca 0.3 )TiO 3 (PDMS/BZT-BCT) and polyvinyl alcohol (PVA). A lead-free BZT-BCT piezoelectric ceramic was prepared via solid-state method and blended into PDMS to form a series of polymer-ceramic composite films, ranging from 5% to 30% by weight. The films were forward/reverse poled with corona poling and their electrical properties were compared to non-poled samples. The H/P-TENG constructed with forward-poled 15 wt% BZT-BCT in PDMS achieved the highest open-circuit voltage, V oc of 127 V, short-circuit current density, J sc of 67 mA m -2 , short-circuit charge density, Q sc of 118 μC m -2 , and peak power density of 7.5 W m -2 , an increase of 190% over pristine PDMS-based TENG. It was discovered that incorporating BZT-BCT into the PDMS matrix improved the triboelectric properties of PDMS. The overlapping electron cloud (OEC) model was used to explain the enhancement and the effect of poling direction of the PDMS/BZT-BCT composite used in H/P-TENG, providing fundamental knowledge of the influence of piezoelectric polarisation on contact electrification., Competing Interests: There are no conflicts to declare., (This journal is © The Royal Society of Chemistry.)

Mechanotransduction by hair cell stereocilia lies at the heart of sound detection in vertebrates. Considerable effort has been put forth to identify proteins that comprise the hair cell mechanotransduction apparatus. TMC1, a member of the transmembrane channel-like (TMC) family, was identified as a core protein of the mechanotransduction complex in hair cells. However, the inability of TMC1 to traffic through the endoplasmic reticulum in heterologous cellular systems has hindered efforts to characterize its function and fully identify its role in mechanotransduction. We developed a novel approach that allowed for the detection of uncharacterized protein regions, which preclude trafficking to the plasma membrane (PM) in heterologous cells. Tagging N-terminal fragments of TMC1 with Aquaporin 3 (AQP3) and GFP fusion reporter, which intrinsically label PM in HEK293 cells, indicated that residues at the edges of amino acid sequence 138-168 invoke intracellular localization and/or degradation. This signal is able to preclude surface localization of PM protein AQP3 in HEK293 cells. Substitutions of the residues by alanine or serine corroborated that the information determining the intracellular retention is present within amino acid sequence 138-168 of TMC1 N-terminus. This novel signal may preclude the proper trafficking of TMC1 to the PM in heterologous cells.

Hearing relies on mechanically gated ion channels present in the actin-rich stereocilia bundles at the apical surface of cochlear hair cells. Our knowledge of the mechanisms underlying the formation and maintenance of the sound-receptive structure is limited. Utilizing a large-scale forward genetic screen in mice, genome mapping and gene complementation tests, we identified Clrn2 as a new deafness gene. The Clrn2 clarinet/clarinet mice (p.Trp4* mutation) exhibit a progressive, early-onset hearing loss, with no overt retinal deficits. Utilizing data from the UK Biobank study, we could show that CLRN2 is involved in human non-syndromic progressive hearing loss. Our in-depth morphological, molecular and functional investigations establish that while it is not required for initial formation of cochlear sensory hair cell stereocilia bundles, clarin-2 is critical for maintaining normal bundle integrity and functioning. In the differentiating hair bundles, lack of clarin-2 leads to loss of mechano-electrical transduction, followed by selective progressive loss of the transducing stereocilia. Together, our findings demonstrate a key role for clarin-2 in mammalian hearing, providing insights into the interplay between mechano-electrical transduction and stereocilia maintenance., (© 2019 The Authors. Published under the terms of the CC BY 4.0 license.)

The pathogenic variant c.144T>G (p.N48K) in the clarin1 gene ( CLRN1 ) results in progressive loss of vision and hearing in Usher syndrome IIIA (USH3A) patients. CLRN1 is predicted to be an essential protein in hair bundles, the mechanosensory structure of hair cells critical for hearing and balance. When expressed in animal models, CLRN1 localizes to the hair bundle, whereas glycosylation-deficient CLRN1 N48K aggregates in the endoplasmic reticulum, with only a fraction reaching the bundle. We hypothesized that the small amount of CLRN1 N48K that reaches the hair bundle does so via an unconventional secretory pathway and that activation of this pathway could be therapeutic. Using genetic and pharmacological approaches, we find that clarin1 knockout ( clrn1 KO/KO ) zebrafish that express the CLRN1 c.144T>G pathogenic variant display progressive hair cell dysfunction, and that CLRN1 N48K is trafficked to the hair bundle via the GRASP55 cargo-dependent unconventional secretory pathway (GCUSP). On expression of GRASP55 mRNA, or on exposure to the drug artemisinin (which activates GCUSP), the localization of CLRN1 N48K to the hair bundles was enhanced. Artemisinin treatment also effectively restored hair cell mechanotransduction and attenuated progressive hair cell dysfunction in clrn1 KO/KO larvae that express CLRN1 c.144T>G , highlighting the potential of artemisinin to prevent sensory loss in CLRN1 c.144T>G patients., Competing Interests: Conflict of interest statement: S.R.G. and K.N.A. have filed a patent application (PCT/US17/55914) titled, “Methods of treating disorders associated with glycosylation defective proteins.”

Cisplatin and other related platinum antineoplastic drugs are commonly used in the treatment of a variety of cancers in both adults and children but are often associated with severe side effects, including hearing loss. Cisplatin's ototoxic effects are multifaceted, culminating in irreversible damage to the mechanosensory hair cells in the inner ear. Platinum drugs act on cancerous cells by forming nuclear DNA adducts, which may initiate signaling leading to cell cycle arrest or apoptosis. Moreover, it was reported that cisplatin may induce mitochondrial DNA damage in non-cancerous cells. Therefore, protecting mitochondria may alleviate cisplatin-induced insult to non-proliferating cells. Thus, it is important to identify agents that shield the mitochondria from cisplatin-induced insult without compromising the anti-tumor actions of the platinum-based drugs. In this study we tested the protective properties of mitochondrial division inhibitor, mdivi-1, a derivative of quinazolinone and a regulator of mitochondrial fission. Interestingly, it has been reported that mdivi-1 increases the apoptosis of cells that are resistant to cisplatin. The ability of mdivi-1 to protect hair cells against cisplatin-induced toxicity was evaluated in a fish model. Wild-type (Tübingen strain), cdh23 mutant, and transgenic pvalb3b ::GFP zebrafish stably expressing GFP in the hair cells were used in this study. Larvae at 5-6 days post fertilization were placed in varying concentrations of cisplatin (50-200 μM) and/or mdivi-1 (1-10 μM) for 16 h. To evaluate hair cell's viability the number of hair bundles per neuromast were counted. To assess hair cell function, we used the FM1-43 uptake assay and recordings of neuromast microphonic potentials. The results showed that mdivi-1 protected hair cells of lateral line neuromasts when they were challenged by 50 μM of cisplatin: viability of hair cells increased almost twice from 19% ± 1.8% to 36% ± 2.0% ( p < 0.001). No protection was observed when higher concentrations of cisplatin were used. In addition, our data were in accord with previously reported results that functional mechanotransduction strongly potentiates cisplatin-induced hair cell toxicity. Together, our results suggest that mitochondrial protection may prevent cisplatin-induced damage to hair cells.

Usher syndrome type III (USH3) characterized by progressive loss of vision and hearing is caused by mutations in the clarin-1 gene (CLRN1). Clrn1 knockout (KO) mice develop hair cell defects by postnatal day 2 (P2) and are deaf by P21-P25. Early onset profound hearing loss in KO mice and lack of information about the cochlear cell type that requires Clrn1 expression pose challenges to therapeutic investigation. We generated KO mice harboring a transgene, TgAC1, consisting of Clrn1-UTR (Clrn1 cDNA including its 5' and 3' UTR) under the control of regulatory elements (Atoh1 3' enhancer/β-globin basal promoter) to direct expression of Clrn1 in hair cells during development and down regulate it postnatally. The KO-TgAC1 mice displayed delayed onset progressive hearing loss associated with deterioration of the hair bundle structure, leading to the hypothesis that hair cell expression of Clrn1 is essential for postnatal preservation of hair cell structure and hearing. Consistent with that hypothesis, perinatal transfection of hair cells in KO-TgAC1 mice with a single injection of AAV-Clrn1-UTR vector showed correlative preservation of the hair bundle structure and hearing through adult life. Further, the efficacy of AAV-Clrn1 vector was significantly attenuated, revealing the potential importance of UTR in gene therapy.

Usher syndrome type III (USH3), characterized by progressive deafness, variable balance disorder and blindness, is caused by destabilizing mutations in the gene encoding the clarin-1 (CLRN1) protein. Here we report a new strategy to mitigate hearing loss associated with a common USH3 mutation CLRN1(N48K) that involves cell-based high-throughput screening of small molecules capable of stabilizing CLRN1(N48K), followed by a secondary screening to eliminate general proteasome inhibitors, and finally an iterative process to optimize structure-activity relationships. This resulted in the identification of BioFocus 844 (BF844). To test the efficacy of BF844, we developed a mouse model that mimicked the progressive hearing loss associated with USH3. BF844 effectively attenuated progressive hearing loss and prevented deafness in this model. Because the CLRN1(N48K) mutation causes both hearing and vision loss, BF844 could in principle prevent both sensory deficiencies in patients with USH3. Moreover, the strategy described here could help identify drugs for other protein-destabilizing monogenic disorders.

Usher syndrome type III (USH3) is characterized by progressive loss of hearing and vision, and varying degrees of vestibular dysfunction. It is caused by mutations that affect the human clarin-1 protein (hCLRN1), a member of the tetraspanin protein family. The missense mutation CLRN1(N48K), which affects a conserved N-glycosylation site in hCLRN1, is a common causative USH3 mutation among Ashkenazi Jews. The affected individuals hear at birth but lose that function over time. Here, we developed an animal model system using zebrafish transgenesis and gene targeting to provide an explanation for this phenotype. Immunolabeling demonstrated that Clrn1 localized to the hair cell bundles (hair bundles). The clrn1 mutants generated by zinc finger nucleases displayed aberrant hair bundle morphology with diminished function. Two transgenic zebrafish that express either hCLRN1 or hCLRN1(N48K) in hair cells were produced to examine the subcellular localization patterns of wild-type and mutant human proteins. hCLRN1 localized to the hair bundles similarly to zebrafish Clrn1; in contrast, hCLRN1(N48K) largely mislocalized to the cell body with a small amount reaching the hair bundle. We propose that this small amount of hCLRN1(N48K) in the hair bundle provides clarin-1-mediated function during the early stages of life; however, the presence of hCLRN1(N48K) in the hair bundle diminishes over time because of intracellular degradation of the mutant protein, leading to progressive loss of hair bundle integrity and hair cell function. These findings and genetic tools provide an understanding and path forward to identify therapies to mitigate hearing loss linked to the CLRN1 mutation., Significance Statement: Mutations in the clarin-1 gene affect eye and ear function in humans. Individuals with the CLRN1(N48K) mutation are born able to hear but lose that function over time. Here, we develop an animal model system using zebrafish transgenesis and gene targeting to provide an explanation for this phenotype. This approach illuminates the role of clarin-1 and the molecular mechanism linked to the CLRN1(N48K) mutation in sensory hair cells of the inner ear. Additionally, the investigation provided an in vivo model to guide future drug discovery to rescue the hCLRN1(N48K) in hair cells., (Copyright © 2015 the authors 0270-6474/15/3510188-14$15.00/0.)

In hair cells of the inner ear, sound or head movement increases tension in fine filaments termed tip links, which in turn convey force to mechanosensitive ion channels to open them. Tip links are formed by a tetramer of two cadherin proteins: protocadherin 15 (PCDH15) and cadherin 23 (CDH23), which have 11 and 27 extracellular cadherin (EC) repeats, respectively. Mutations in either protein cause inner ear disorders in mice and humans. We showed recently that these two cadherins bind tip-to-tip in a "handshake" mode that involves the EC1 and EC2 repeats of both proteins. However, a paucity of appropriate animal models has slowed our understanding both of the interaction and of how mutations of residues within the predicted interface compromise tip link integrity. Here, we present noddy, a new mouse model for hereditary deafness. Identified in a forward genetic screen, noddy homozygotes lack inner ear function. Mapping and sequencing showed that noddy mutant mice harbor an isoleucine-to-asparagine (I108N) mutation in the EC1 repeat of PCDH15. Residue I108 interacts with CDH23 EC2 in the handshake and its mutation impairs the interaction in vitro. The noddy mutation allowed us to determine the consequences of blocking the handshake in vivo: tip link formation and bundle morphology are disrupted, and mechanotransduction channels fail to remain open at rest. These results offer new insights into the interaction between PCDH15 and CDH23 and help explain the etiology of human deafness linked to mutations in the tip-link interface.

Cell immobilization is one of the common techniques for increasing the overall cell concentration and productivity. Bacillus amyloliquefaciens MBL27 cells were immobilized in calcium alginate beads and it is a promising method for repeated AMP (antimicrobial protein) production. The present study aimed at determining the optimal conditions for immobilization of B. amyloliquefaciens MBL27 cells in calcium alginate beads and the operational stability for enhanced production of the AMP. AMP production with free and immobilized cells was also done. In batch fermentation, maximum AMP production (7300 AU (arbitrary units)/ml against Staphylococcus aureus) was obtained with immobilized cells in shake flasks under optimized parameters such as 3% (w/v) sodium alginate, 136 mM CaCl2 with 350 alginate beads/flask of 2.7-3.0 mm diameter. In repeated cultivation, the highest activity was obtained after the second cycle of use and approx. 94% production was noted up to the fifth cycle. The immobilized cells of B. amyloliquefaciens MBL27 in alginate beads are more efficient for the production of AMP and had good stability. The potential application of AMP as a wound healant and the need for development of economical methods for improved production make whole cell immobilization an excellent alternative method for enhanced AMP production.