Custom Loss Functions in XGBoost Algorithm for Enhanced Critical Error Mitigation in Drill-Wear Analysis of Melamine-Faced Chipboard.

The advancement of machine learning in industrial applications has necessitated the development of tailored solutions to address specific challenges, particularly in multi-class classification tasks. This study delves into the customization of loss functions within the eXtreme Gradient Boosting (XGBoost) algorithm, which is a critical step in enhancing the algorithm's performance for specific applications. Our research is motivated by the need for precision and efficiency in the industrial domain, where the implications of misclassification can be substantial. We focus on the drill-wear analysis of melamine-faced chipboard, a common material in furniture production, to demonstrate the impact of custom loss functions. The paper explores several variants of Weighted Softmax Loss Functions, including Edge Penalty and Adaptive Weighted Softmax Loss, to address the challenges of class imbalance and the heightened importance of accurately classifying edge classes. Our findings reveal that these custom loss functions significantly reduce critical errors in classification without compromising the overall accuracy of the model. This research not only contributes to the field of industrial machine learning by providing a nuanced approach to loss function customization but also underscores the importance of context-specific adaptations in machine learning algorithms. The results showcase the potential of tailored loss functions in balancing precision and efficiency, ensuring reliable and effective machine learning solutions in industrial settings.

Advanced Feature Extraction Methods from Images of Drillings in Melamine Faced Chipboard for Automatic Diagnosis of Drill Wear.

In this paper, a novel approach to evaluation of feature extraction methodologies is presented. In the case of machine learning algorithms, extracting and using the most efficient features is one of the key problems that can significantly influence overall performance. It is especially the case with parameter-heavy problems, such as tool condition monitoring. In the presented case, images of drilled holes are considered, where state of the edge and the overall size of imperfections have high influence on product quality. Finding and using a set of features that accurately describes the differences between the edge that is acceptable or too damaged is not always straightforward. The presented approach focuses on detailed evaluation of various feature extraction approaches. Each chosen method produced a set of features, which was then used to train a selected set of classifiers. Five initial feature sets were obtained, and additional ones were derived from them. Different voting methods were used for ensemble approaches. In total, 38 versions of the classifiers were created and evaluated. Best accuracy was obtained by the ensemble approach based on Weighted Voting methodology. A significant difference was shown between different feature extraction methods, with a total difference of 11.14% between the worst and best feature set, as well as a further 0.2% improvement achieved by using the best voting approach.

Improved Drill State Recognition during Milling Process Using Artificial Intelligence.

In this article, an automated method for tool condition monitoring is presented. When producing items in large quantities, pointing out the exact time when the element needs to be exchanged is crucial. If performed too early, the operator gets rid of a good drill, also resulting in production downtime increase if this operation is repeated too often. On the other hand, continuing production with a worn tool might result in a poor-quality product and financial loss for the manufacturer. In the presented approach, drill wear is classified using three states representing decreasing quality: green, yellow and red. A series of signals were collected as training data for the classification algorithms. Measurements were saved in separate data sets with corresponding time windows. A total of ten methods were evaluated in terms of overall accuracy and the number of misclassification errors. Three solutions obtained an acceptable accuracy rate above 85%. Algorithms were able to assign states without the most undesirable red-green and green-red errors. The best results were achieved by the Extreme Gradient Boosting algorithm. This approach achieved an overall accuracy of 93.33%, and the only misclassification was the yellow sample assigned as green. The presented solution achieves good results and can be applied in industry applications related to tool condition monitoring.

Staphylococcal saoABC Operon Codes for a DNA-Binding Protein SaoC Implicated in the Response to Nutrient Deficit.

Whilst a large number of regulatory mechanisms for gene expression have been characterised to date, transcription regulation in bacteria still remains an open subject. In clinically relevant and opportunistic pathogens, such as Staphylococcus aureus, transcription regulation is of great importance for host-pathogen interactions. In our study we investigated an operon, exclusive to staphylococci, that we name saoABC. We showed that SaoC binds to a conserved sequence motif present upstream of the saoC gene, which likely provides a negative feedback loop. We have also demonstrated that S. aureus ΔsaoB and ΔsaoC mutants display altered growth dynamics in non-optimal media; ΔsaoC exhibits decreased intracellular survival in human dermal fibroblasts, whereas ΔsaoB produces an elevated number of persisters, which is also elicited by inducible production of SaoC in ΔsaoBΔsaoC double mutant. Moreover, we have observed changes in the expression of saoABC operon genes during either depletion of the preferential carbon or the amino acid source as well as during acidification. Comparative RNA-Seq of the wild type and ΔsaoC mutant demonstrated that SaoC influences transcription of genes involved in amino acid transport and metabolism, and notably of those coding for virulence factors. Our results suggest compellingly that saoABC operon codes for a DNA-binding protein SaoC, a novel staphylococcal transcription factor, and its antagonist SaoB. We linked SaoC to the response to nutrient deficiency, a stress that has a great impact on host-pathogen interactions. That impact manifests in SaoC influence on persister formation and survival during internalisation to host cells, as well as on the expression of genes of virulence factors that may potentially result in profound alternations in the pathogenic phenotype. Investigation of such novel regulatory mechanisms is crucial for our understanding of the dynamics of interactions between pathogenic bacteria and host cells, particularly in the case of clinically relevant, opportunistic pathogens such as Staphylococcus aureus.

Towards understanding the novel adhesin function of Candida albicans phosphoglycerate mutase at the pathogen cell surface: some structural analysis of the interactions with human host extracellular matrix proteins.

Although many atypical proteinaceous cell wall components that belong to a group of multitasking, "moonlighting" proteins, have been repeatedly identified in numerous pathogenic microorganisms, their novel extracellular functions and secretion mechanisms remain largely unrecognized. In Candida albicans, one of the most common fungal pathogens in humans, phosphoglycerate mutase (Gpm1) - a cytoplasmic enzyme involved in the glycolysis pathway - has been shown to occur on the cell surface and has been identified as a potentially important virulence factor. In this study, we demonstrated tight binding of C. albicans Gpm1 to the candidal cell surface, thus suggesting that the readsorption of soluble Gpm1 from the external environment could be a likely mechanism leading to the presence of this moonlighting protein on the pathogen surface. Several putative Gpm1-binding receptors on the yeast surface were identified. The affinities of Gpm1 to human vitronectin (VTR) and fibronectin (FN) were characterized with surface plasmon resonance measurements, and the dissociation constants of the complexes formed were determined to be in the order of 10-8 M. The internal Gpm1 sequence motifs, directly interacting with VTR (aa 116-158) and FN (aa 138-175) were mapped using chemical crosslinking and mass spectrometry. Synthetic peptides with matching sequences significantly inhibited formation of the Gpm1-VTR and Gpm1-FN complexes. A molecular model of the Gpm1-VTR complex was developed. These results provide the first structural insights into the adhesin function of candidal surface-exposed Gpm1.

Decision Confidence Assessment in Multi-Class Classification.

This paper presents a novel approach to the assessment of decision confidence when multi-class recognition is concerned. When many classification problems are considered, while eliminating human interaction with the system might be one goal, it is not the only possible option-lessening the workload of human experts can also bring huge improvement to the production process. The presented approach focuses on providing a tool that will significantly decrease the amount of work that the human expert needs to conduct while evaluating different samples. Instead of hard classification, which assigns a single label to each class, the described solution focuses on evaluating each case in terms of decision confidence-checking how sure the classifier is in the case of the currently processed example, and deciding if the final classification should be performed, or if the sample should instead be manually evaluated by a human expert. The method can be easily adjusted to any number of classes. It can also focus either on the classification accuracy or coverage of the used dataset, depending on user preferences. Different confidence functions are evaluated in that aspect. The results obtained during experiments meet the initial criteria, providing an acceptable quality for the final solution.

Human skin microbiota-friendly lysostaphin.

Growing antibiotic resistance of bacteria is a burning problem of human and veterinary medicine. Expansion and introduction of novel microbicidal therapeutics is highly desirable. However, antibiotic treatment disturbs the balance of physiological microbiota by changing its qualitative and/or quantitative composition, resulting in a number of adverse effects that include secondary infections. Although such dysbiosis may be reversed by the treatment with probiotics, a more attractive alternative is the use of antibiotics that target only pathogens, while sparing the commensals. Here, we describe lysostaphin LSp222, an enzyme produced naturally by Staphylococcus pseudintermedius 222. LSp222 is highly effective against S. aureus, including its multi-drug resistant strains. Importantly, the inhibitory concentration for S. epidermidis, the predominant commensal in healthy human skin, is at least two orders of magnitude higher compared to S. aureus. Such significant therapeutic window makes LSp222 a microbiota-friendly antibacterial agent with a potential application in the treatment of S. aureus-driven skin infections.

Isolation, Identification, and Bioinformatic Analysis of Antibacterial Proteins and Peptides from Immunized Hemolymph of Red Palm Weevil Rhynchophorus ferrugineus.

Red palm weevil (Rhynchophorus ferrugineus Olivier, 1791, Coleoptera: Curculionidae) is a destructive pest of palms, rapidly extending its native geographical range and causing large economic losses worldwide. The present work describes isolation, identification, and bioinformatic analysis of antibacterial proteins and peptides from the immunized hemolymph of this beetle. In total, 17 different bactericidal or bacteriostatic compounds were isolated via a series of high-pressure liquid chromatography steps, and their partial amino acid sequences were determined by N-terminal sequencing or by mass spectrometry. The bioinformatic analysis of the results facilitated identification and description of corresponding nucleotide coding sequences for each peptide and protein, based on the recently published R. ferrugineus transcriptome database. The identified compounds are represented by several well-known bactericidal factors: two peptides similar to defensins, one cecropin-A1-like peptide, and one attacin-B-like protein. Interestingly, we have also identified some unexpected compounds comprising five isoforms of pheromone-binding proteins as well as seven isoforms of odorant-binding proteins. The particular role of these factors in insect response to bacterial infection needs further investigation.

Application of Siamese Networks to the Recognition of the Drill Wear State Based on Images of Drilled Holes.

In this article, a Siamese network is applied to the drill wear classification problem. For furniture companies, one of the main problems that occurs during the production process is finding the exact moment when the drill should be replaced. When the drill is not sharp enough, it can result in a poor quality product and therefore generate some financial loss for the company. In various approaches to this problem, usually, three classes are considered: green for a drill that is sharp, red for the opposite, and yellow for a tool that is suspected of being worn out, requiring additional evaluation by a human expert. In the above problem, it is especially important that the green and the red classes not be mistaken, since such errors have the highest probability of generating financial loss for the manufacturer. Most of the solutions analysing this problem are too complex, requiring specialized equipment, high financial investment, or both, without guaranteeing that the obtained results will be satisfactory. In the approach presented in this paper, images of drilled holes are used as the training data for the Siamese network. The presented solution is much simpler in terms of the data collection methodology, does not require a large financial investment for the initial equipment, and can accurately qualify drill wear based on the chosen input. It also takes into consideration additional manufacturer requirements, like no green-red misclassifications, that are usually omitted in existing solutions.

Structural Insights into the Interactions of Candidal Enolase with Human Vitronectin, Fibronectin and Plasminogen.

Significant amounts of enolase-a cytosolic enzyme involved in the glycolysis pathway-are exposed on the cell surface of Candida yeast. It has been hypothesized that this exposed enolase form contributes to infection-related phenomena such as fungal adhesion to human tissues, and the activation of fibrinolysis and extracellular matrix degradation. The aim of the present study was to characterize, in structural terms, the protein-protein interactions underlying these moonlighting functions of enolase. The tight binding of human vitronectin, fibronectin and plasminogen by purified C. albicans and C. tropicalis enolases was quantitatively analyzed by surface plasmon resonance measurements, and the dissociation constants of the formed complexes were determined to be in the 10-7-10-8 M range. In contrast, the binding of human proteins by the S.cerevisiae enzyme was much weaker. The chemical cross-linking method was used to map the sites on enolase molecules that come into direct contact with human proteins. An internal motif 235DKAGYKGKVGIAMDVASSEFYKDGK259 in C. albicans enolase was suggested to contribute to the binding of all three human proteins tested. Models for these interactions were developed and revealed the sites on the enolase molecule that bind human proteins, extensively overlap for these ligands, and are well-separated from the catalytic activity center.

Chromosomal localization of PemIK toxin-antitoxin system results in the loss of toxicity - Characterization of pemIKSa1-Sp from Staphylococcus pseudintermedius.

Toxin-antitoxin (TA) systems are ubiquitous in bacteria and on numerous occasions have been postulated to play a role in virulence of pathogens. Some Staphylococcus aureus strains carry a plasmid, which encodes the highly toxic PemIKSa TA system involved in maintenance of the plasmid but also implicated in modulation of gene expression. Here we showed that pemIKSa1-Sp TA system, homologous to the plasmid-encoded PemIKSa, is present in virtually each chromosome of S. pseudintermedius strain, however exhibits sequence heterogeneity. This results in two length variants of the PemKSa1-Sp toxin. The shorter (96 aa), C-terminally truncated toxin is enzymatically inactive, whereas the full length (112 aa) variant is an RNase, though nontoxic to the host cells. The lack of toxicity of the active PemKSa-Sp2 toxin is explained by increased substrate specificity. The pemISa1-Sp antitoxin gene seems pseudogenized, however, the whole pemIKSa1-Sp system is transcriptionally active. When production of N-terminally truncated antitoxins using alternative start codons is assumed, there are five possible length variants. Here we showed that even substantially truncated antitoxins are able to interact with PemKSa-Sp2 toxin and inhibit its RNase activity. Moreover, the antitoxins can rescue bacterial cells from toxic effects of overexpression of plasmid-encoded PemKSa toxin. Collectively, our data indicates that, contrary to the toxic plasmid-encoded PemIKSa TA system, location of pemIKSa1-Sp in the chromosome of S. pseudintermedius results in the loss of its toxicity. Interestingly, the retained RNase activity of PemKSa1-Sp2 toxin and functionality of the putative, N-terminally truncated antitoxins suggest the existence of evolutionary pressure for alleviation/mitigation of the toxin's toxicity and retention of the inhibitory activity of the antitoxin, respectively.

Prevalence of Antibiotic and Heavy Metal Resistance Determinants and Virulence-Related Genetic Elements in Plasmids of Staphylococcus aureus.

The use of antibiotics on a mass scale, particularly in farming, and their release into the environment has led to a rapid emergence of resistant bacteria. Once emerged, resistance determinants are spread by horizontal gene transfer among strains of the same as well as disparate bacterial species. Their accumulation in free-living as well as livestock and community-associated strains results in the widespread multiple-drug resistance among clinically relevant species posing an increasingly pressing problem in healthcare. One of these clinically relevant species is Staphylococcus aureus, a common cause of hospital and community outbreaks. Among the rich diversity of mobile genetic elements regularly occurring in S. aureus such as phages, pathogenicity islands, and staphylococcal cassette chromosomes, plasmids are the major mean for dissemination of resistance determinants and virulence factors. Unfortunately, a vast number of whole-genome sequencing projects does not aim for complete sequence determination, which results in a disproportionately low number of known complete plasmid sequences. To address this problem we determined complete plasmid sequences derived from 18 poultry S. aureus strains and analyzed the prevalence of antibiotic and heavy metal resistance determinants, genes of virulence factors, as well as genetic elements relevant for their maintenance. Some of the plasmids have been reported before and are being found in clinical isolates of strains typical for humans or human ones of livestock origin. This shows that livestock-associated staphylococci are a significant reservoir of resistance determinants and virulence factors. Nevertheless, nearly half of the plasmids were unknown to date. In this group we found a potentially mobilizable plasmid pPA3 being a unique example of accumulation of resistance determinants and virulence factors likely stabilized by a presence of a toxin-antitoxin system.

Joint Genomic and Proteomic Analysis Identifies Meta-Trait Characteristics of Virulent and Non-virulent Staphylococcus aureus Strains.

Staphylococcus aureus is an opportunistic pathogen of humans and warm-blooded animals and presents a growing threat in terms of multi-drug resistance. Despite numerous studies, the basis of staphylococcal virulence and switching between commensal and pathogenic phenotypes is not fully understood. Using genomics, we show here that S. aureus strains exhibiting virulent (VIR) and non-virulent (NVIR) phenotypes in a chicken embryo infection model genetically fall into two separate groups, with the VIR group being much more cohesive than the NVIR group. Significantly, the genes encoding known staphylococcal virulence factors, such as clumping factors, are either found in different allelic variants in the genomes of NVIR strains (compared to VIR strains) or are inactive pseudogenes. Moreover, the pyruvate carboxylase and gamma-aminobutyrate permease genes, which were previously linked with virulence, are pseudogenized in NVIR strain ch22. Further, we use comprehensive proteomics tools to characterize strains that show opposing phenotypes in a chicken embryo virulence model. VIR strain CH21 had an elevated level of diapolycopene oxygenase involved in staphyloxanthin production (protection against free radicals) and expressed a higher level of immunoglobulin-binding protein Sbi on its surface compared to NVIR strain ch22. Furthermore, joint genomic and proteomic approaches linked the elevated production of superoxide dismutase and DNA-binding protein by NVIR strain ch22 with gene duplications.

Identification of Antioxidant Peptides in Enzymatic Hydrolysates of Carp (Cyprinus Carpio) Skin Gelatin.

The protein by-products from carp (Cyprinus carpio) are normally discarded as industrial waste during fish processing. The objective of this study was to identify and characterise the peptides with a potential antioxidant activity that are released from carp skin proteins during hydrolysis by the Protamex enzyme mixture. This study shows that a hydrolysate of carp skin gelatin and its reversed-phase chromatography fractions have strong in vitro antioxidant properties. Among these fractions, the alanine-tyrosine (Ala-Tyr) dipeptide was identified as the major compound with high antioxidant potential. The peptide has good stability during in vitro enzymatic digestion assay and can inhibit the angiotensin-converting enzyme (ACE). In conclusion, our study proves that both the unfractionated hydrolysate of carp skin gelatin and the above-mentioned Ala-Tyr dipeptide represents attractive novel compounds for the formulation of antioxidant foods.

Identification of novel mazEF/pemIK family toxin-antitoxin loci and their distribution in the Staphylococcus genus.

The versatile roles of toxin-antitoxin (TA) systems in bacterial physiology and pathogenesis have been investigated for more than three decades. Diverse TA loci in Bacteria and Archaea have been identified in genome-wide studies. The advent of massive parallel sequencing has substantially expanded the number of known bacterial genomic sequences over the last 5 years. In staphylococci, this has translated into an impressive increase from a few tens to a several thousands of available genomes, which has allowed us for the re-evalution of prior conclusions. In this study, we analysed the distribution of mazEF/pemIK family TA system operons in available staphylococcal genomes and their prevalence in mobile genetic elements. 10 novel m azEF/pemIK homologues were identified, each with a corresponding toxin that plays a potentially different and undetermined physiological role. A detailed characterisation of these TA systems would be exceptionally useful. Of particular interest are those associated with an SCCmec mobile genetic element (responsible for multidrug resistance transmission) or representing the joint horizontal transfer of TA systems and determinants of vancomycin resistance from enterococci. The involvement of TA systems in maintaining mobile genetic elements and the associations between novel mazEF/pemIK loci and those which carry drug resistance genes highlight their potential medical importance.

Identification of Secreted Exoproteome Fingerprints of Highly-Virulent and Non-Virulent Staphylococcus aureus Strains.

Staphylococcus aureus is a commensal inhabitant of skin and mucous membranes in nose vestibule but also an important opportunistic pathogen of humans and livestock. The extracellular proteome as a whole constitutes its major virulence determinant; however, the involvement of particular proteins is still relatively poorly understood. In this study, we compared the extracellular proteomes of poultry-derived S. aureus strains exhibiting a virulent (VIR) and non-virulent (NVIR) phenotype in a chicken embryo experimental infection model with the aim to identify proteomic signatures associated with the particular phenotypes. Despite significant heterogeneity within the analyzed proteomes, we identified alpha-haemolysin and bifunctional autolysin as indicators of virulence, whereas glutamylendopeptidase production was characteristic for non-virulent strains. Staphopain C (StpC) was identified in both the VIR and NVIR proteomes and the latter fact contradicted previous findings suggesting its involvement in virulence. By supplementing NVIR, StpC-negative strains with StpC, and comparing the virulence of parental and supplemented strains, we demonstrated that staphopain C alone does not affect staphylococcal virulence in a chicken embryo model.

Daptomycin-resistant Staphylococcus pettenkoferi of human origin.

The importance of nosocomial infections caused by coagulase-negative staphylococci is constantly growing. The threat primarily affects immunocompromised patients, the elderly and neonates, particularly after invasive surgery. The problem is fundamentally exacerbated by expanding antibacterial drug resistance. A case report is presented of an 86-year-old patient who underwent a ruptured abdominal aortic aneurysm surgery and developed septicaemia upon surgical wound infection. The causal agent was likely a carbapenem-resistant Klebsiella pneumoniae, however, daptomycin-resistant Staphylococcus pettenkoferi was identified in blood cultures in the absence of daptomycin treatment. To the authors' knowledge, the case study presented is the first published episode of daptomycin-resistant S. pettenkoferi strain.

A peptide factor secreted by Staphylococcus pseudintermedius exhibits properties of both bacteriocins and virulence factors.

Staphylococcus pseudintermedius is a common commensal bacterium colonizing the skin and mucosal surfaces of household animals. However, it has recently emerged as a dangerous opportunistic pathogen, comparable to S. aureus for humans. The epidemiological situation is further complicated by the increasing number of methicillin-resistant S. pseudintermedius infections and evidence of gene transmission driving antibiotic resistance between staphylococci colonizing human and zoonotic hosts. In the present study, we describe a unique peptide, BacSp222, that possesses features characteristic of both bacteriocins and virulence factors. BacSp222 is secreted in high quantities by S. pseudintermedius strain 222 isolated from dog skin lesions. This linear, fifty-amino-acid highly cationic peptide is plasmid-encoded and does not exhibit significant sequence similarities to any other known peptides or proteins. BacSp222 kills gram-positive bacteria (at doses ranging from 0.1 to several micromol/l) but also demonstrates significant cytotoxic activities towards eukaryotic cells at slightly higher concentrations. Moreover, at nanomolar concentrations, the peptide also possesses modulatory properties, efficiently enhancing interferon gamma-induced nitric oxide release in murine macrophage-like cell lines. BacSp222 appears to be one of the first examples of multifunctional peptides that breaks the convention of splitting bacteriocins and virulence factors into two unrelated groups.

A systematic investigation of the stability of green fluorescent protein fusion proteins.

X-ray crystallography provides important insights into structure-function relationship in biomolecules. However, protein crystals are usually hard to obtain which hinders our understanding of multiple important processes. Crystallization requires large amount of protein sample, whereas recombinant proteins are often unstable or insoluble. Green fluorescent protein (GFP) fusion is one of the approaches to increase protein synthesis, solubility and stability, facilitating crystallization. In this study we analyze the influence of the linker length, composition and the position of GFP relative to the fusion partner on the fusion protein production and stability. To this end, multiple constructs of enzymatically impaired variant of PemKSa toxin from Staphylococcus aureus CH91 fused to GFP were generated. Fusion protein production in Escherichia coli was evaluated. The proteins were purified and their stability tested. PemKSa-α14aa-GFP fusion provided best production and stability. Obtained results demonstrate the importance of optimization of fusion protein construct, including linker selection and the order of fusion partners, in obtaining high quantities of stable protein for crystallization.

Species determination within Staphylococcus genus by extended PCR-restriction fragment length polymorphism of saoC gene.

Genetic methods based on PCR-restriction fragment length polymorphism (RFLP) are widely used for microbial species determination. In this study, we present the application of saoC gene as an effective tool for species determination and within-species diversity analysis for Staphylococcus genus. The unique sequence diversity of saoC allows us to apply four restriction enzymes to obtain RFLP patterns, which appear highly distinctive even among closely related species as well as atypical isolates of environmental origin. Such patterns were successfully obtained for 26 species belonging to Staphylococcus genus. What is more, tracing polymorphisms detected by different restriction enzymes allowed for basic phylogeny analysis for Staphylococcus aureus, which is potentially applicable for other staphylococcal species.