MLm5C: A high-precision human RNA 5-methylcytosine sites predictor based on a combination of hybrid machine learning models.

RNA modification serves as a pivotal component in numerous biological processes. Among the prevalent modifications, 5-methylcytosine (m5C) significantly influences mRNA export, translation efficiency and cell differentiation and are also associated with human diseases, including Alzheimer's disease, autoimmune disease, cancer, and cardiovascular diseases. Identification of m5C is critically responsible for understanding the RNA modification mechanisms and the epigenetic regulation of associated diseases. However, the large-scale experimental identification of m5C present significant challenges due to labor intensity and time requirements. Several computational tools, using machine learning, have been developed to supplement experimental methods, but identifying these sites lack accuracy and efficiency. In this study, we introduce a new predictor, MLm5C, for precise prediction of m5C sites using sequence data. Briefly, we evaluated eleven RNA sequence-derived features with four basic machine learning algorithms to generate baseline models. From these 44 models, we ranked them based on their performance and subsequently stacked the Top 20 baseline models as the best model, named MLm5C. The MLm5C outperformed the-state-of-the-art predictors. Notably, the optimization of the sequence length surrounding the modification sites significantly improved the prediction performance. MLm5C is an invaluable tool in accelerating the detection of m5C sites within the human genome, thereby facilitating in the characterization of their roles in post-transcriptional regulation.

Elucidating Key Characteristics of PFAS Binding to Human Peroxisome Proliferator-Activated Receptor Alpha: An Explainable Machine Learning Approach.

Per- and polyfluoroalkyl substances (PFAS) are widely employed anthropogenic fluorinated chemicals known to disrupt hepatic lipid metabolism by binding to human peroxisome proliferator-activated receptor alpha (PPARα). Therefore, screening for PFAS that bind to PPARα is of critical importance. Machine learning approaches are promising techniques for rapid screening of PFAS. However, traditional machine learning approaches lack interpretability, posing challenges in investigating the relationship between molecular descriptors and PPARα binding. In this study, we aimed to develop a novel, explainable machine learning approach to rapidly screen for PFAS that bind to PPARα. We calculated the PPARα-PFAS binding score and 206 molecular descriptors for PFAS. Through systematic and objective selection of important molecular descriptors, we developed a machine learning model with good predictive performance using only three descriptors. The molecular size (b_single) and electrostatic properties (BCUT_PEOE_3 and PEOE_VSA_PPOS) are important for PPARα-PFAS binding. Alternative PFAS are considered safer than their legacy predecessors. However, we found that alternative PFAS with many carbon atoms and ether groups exhibited a higher affinity for PPARα. Therefore, confirming the toxicity of these alternative PFAS compounds with such characteristics through biological experiments is important.

Automatic Generation of SBML Kinetic Models from Natural Language Texts Using GPT.

Kinetic modeling is an essential tool in systems biology research, enabling the quantitative analysis of biological systems and predicting their behavior. However, the development of kinetic models is a complex and time-consuming process. In this article, we propose a novel approach called KinModGPT, which generates kinetic models directly from natural language text. KinModGPT employs GPT as a natural language interpreter and Tellurium as an SBML generator. We demonstrate the effectiveness of KinModGPT in creating SBML kinetic models from complex natural language descriptions of biochemical reactions. KinModGPT successfully generates valid SBML models from a range of natural language model descriptions of metabolic pathways, protein-protein interaction networks, and heat shock response. This article demonstrates the potential of KinModGPT in kinetic modeling automation.

MLAGO: machine learning-aided global optimization for Michaelis constant estimation of kinetic modeling.

BACKGROUND: Kinetic modeling is a powerful tool for understanding the dynamic behavior of biochemical systems. For kinetic modeling, determination of a number of kinetic parameters, such as the Michaelis constant (Km), is necessary, and global optimization algorithms have long been used for parameter estimation. However, the conventional global optimization approach has three problems: (i) It is computationally demanding. (ii) It often yields unrealistic parameter values because it simply seeks a better model fitting to experimentally observed behaviors. (iii) It has difficulty in identifying a unique solution because multiple parameter sets can allow a kinetic model to fit experimental data equally well (the non-identifiability problem). RESULTS: To solve these problems, we propose the Machine Learning-Aided Global Optimization (MLAGO) method for Km estimation of kinetic modeling. First, we use a machine learning-based Km predictor based only on three factors: EC number, KEGG Compound ID, and Organism ID, then conduct a constrained global optimization-based parameter estimation by using the machine learning-predicted Km values as the reference values. The machine learning model achieved relatively good prediction scores: RMSE = 0.795 and R2 = 0.536, making the subsequent global optimization easy and practical. The MLAGO approach reduced the error between simulation and experimental data while keeping Km values close to the machine learning-predicted values. As a result, the MLAGO approach successfully estimated Km values with less computational cost than the conventional method. Moreover, the MLAGO approach uniquely estimated Km values, which were close to the measured values. CONCLUSIONS: MLAGO overcomes the major problems in parameter estimation, accelerates kinetic modeling, and thus ultimately leads to better understanding of complex cellular systems. The web application for our machine learning-based Km predictor is accessible at https://sites.google.com/view/kazuhiro-maeda/software-tools-web-apps , which helps modelers perform MLAGO on their own parameter estimation tasks.

Inflammation and Bone Metabolism in Rheumatoid Arthritis: Molecular Mechanisms of Joint Destruction and Pharmacological Treatments.

Rheumatoid arthritis (RA) is an inflammatory disease characterized by a variety of symptoms and pathologies often presenting with polyarthritis. The primary symptom in the initial stage is joint swelling due to synovitis. With disease progression, cartilage and bone are affected to cause joint deformities. Advanced osteoarticular destruction and deformation can cause irreversible physical disabilities. Physical disabilities not only deteriorate patients' quality of life but also have substantial medical economic effects on society. Therefore, prevention of the progression of osteoarticular destruction and deformation is an important task. Recent studies have progressively improved our understanding of the molecular mechanism by which synovitis caused by immune disorders results in activation of osteoclasts; activated osteoclasts in turn cause bone destruction and para-articular osteoporosis. In this paper, we review the mechanisms of bone metabolism under physiological and RA conditions, and we describe the effects of therapeutic intervention against RA on bone.

The Regulation of Bone Metabolism and Disorders by Wnt Signaling.

Wnt, a secreted glycoprotein, has an approximate molecular weight of 40 kDa, and it is a cytokine involved in various biological phenomena including ontogeny, morphogenesis, carcinogenesis, and maintenance of stem cells. The Wnt signaling pathway can be classified into two main pathways: canonical and non-canonical. Of these, the canonical Wnt signaling pathway promotes osteogenesis. Sclerostin produced by osteocytes is an inhibitor of this pathway, thereby inhibiting osteogenesis. Recently, osteoporosis treatment using an anti-sclerostin therapy has been introduced. In this review, the basics of Wnt signaling, its role in bone metabolism and its involvement in skeletal disorders have been covered. Furthermore, the clinical significance and future scopes of Wnt signaling in osteoporosis, osteoarthritis, rheumatoid arthritis and neoplasia are discussed.

Long negative feedback loop enhances period tunability of biological oscillators.

Oscillatory phenomena play a major role in organisms. In some biological oscillations such as cell cycles and heartbeats, the period can be tuned without significant changes in the amplitude. This property is called (period) tunability, one of the prominent features of biological oscillations. However, how biological oscillators produce tunable oscillations remains largely unexplored. We tackle this question using computational experiments. It has been reported that positive-plus-negative feedback oscillators produce tunable oscillations through the hysteresis-based mechanism. First, in this study, we confirmed that positive-plus-negative feedback oscillators generate tunable oscillations. Second, we found that tunability is positively correlated with the dynamic range of oscillations. Third, we showed that long negative feedback oscillators without any additional positive feedback loops can produce tunable oscillations. Finally, we computationally demonstrated that by lengthening the negative feedback loop, the Repressilator, known as a non-tunable synthetic gene oscillator, can be converted into a tunable oscillator. This work provides synthetic biologists with clues to design tunable gene oscillators.

Postotic and preotic cranial neural crest cells differently contribute to thyroid development.

Thyroid development and formation vary among species, but in most species the thyroid morphogenesis consists of five stages: specification, budding, descent, bilobation and folliculogenesis. The detailed mechanisms of these stages have not been fully clarified. During early development, the cranial neural crest (CNC) contributes to the thyroid gland. The removal of the postotic CNC (corresponding to rhombomeres 6, 7 and 8, also known as the cardiac neural crest) results in abnormalities of the cardiovascular system, thymus, parathyroid glands, and thyroid gland. To investigate the influence of the CNC on thyroid bilobation process, we divided the CNC into two regions, the postotic CNC and the preotic CNC (from the mesencephalon to rhombomere 5) regions and examined. We found that preotic CNC-ablated embryos had a unilateral thyroid lobe, and confirmed the presence of a single lobe or the absence of lobes in postotic CNC-ablated chick embryos. The thyroid anlage in each region-ablated embryos was of a normal size at the descent stage, but at a later stage, the thyroid in preotic CNC-ablated embryos was of a normal size, conflicting with a previous report in which the thyroid was reduced in size in the postotic CNC-ablated embryos. The postotic CNC cells differentiated into connective tissues of the thyroid in quail-to-chick chimeras. In contrast, the preotic CNC cells did not differentiate into connective tissues of the thyroid. We found that preotic CNC cells encompassed the thyroid anlage from the specification stage to the descent stage. Finally, we found that endothelin-1 and endothelin type A receptor-knockout mice and bosentan (endothelin receptor antagonist)-treated chick embryos showed bilobation anomalies that included single-lobe formation. Therefore, not only the postotic CNC, but also the preotic CNC plays an important role in thyroid morphogenesis.

BK-UM in patients with recurrent ovarian cancer or peritoneal cancer: a first-in-human phase-I study.

BACKGROUND: BK-UM (CRM197) is a mutant form of diphtheria toxin and a specific inhibitor of heparin-binding epidermal growth factor-like growth factor (HB-EGF). We assessed the safety, pharmacokinetics, recommended dose, and efficacy of BK-UM in patients with recurrent ovarian cancer (OC) or peritoneal cancer (PC), and measured HB-EGF levels in serum and abdominal fluid after BK-UM administration. METHODS: Eleven patients with advanced or recurrent OC or PC were enrolled and treated with BK-UM via the intraperitoneal route. The dose was escalated (1.0, 2.0, 3.3, and 5.0 mg/m2) using a 3 + 3 design. RESULTS: Eight of 11 patients completed treatment. No dose-limiting toxicity (DLT) was experienced at dose levels 1 (1.0 mg/m2) and 2 (2.0 mg/m2). Grade 3 transient hypotension as an adverse event (defined as a DLT in the present study) was observed in two of four patients at dose level 3 (3.3 mg/m2). Treatment with BK-UM was associated with decreases in HB-EGF levels in serum and abdominal fluid in seven of 11 patients and five of eight patients, respectively. Clinical outcomes included a partial response in one patient, stable disease in five patients, and progressive disease in five patients. CONCLUSIONS: BK-UM was well tolerated at doses of 1.0 and 2.0 mg/m2, with evidence for clinical efficacy in patients with recurrent OC or PC. A dose of 2.0 mg/m2 BK-UM is recommended for subsequent clinical trials. TRIAL REGISTRATION: This trial was prospectively performed as an investigator-initiated clinical trial. The trial numbers are UMIN000001002 and UMIN000001001, with registration dates of 1/30/2008 and 2/4/2008, respectively. UMIN000001001 was registered as a trial for the continuous administration of BK-UM after UMIN000001002 .

Development of selective medium for IMP-type carbapenemase-producing Enterobacteriaceae in stool specimens.

BACKGROUND: Identification of carbapenemase-producing Enterobacteriaceae (CPE) in faecal specimens is challenging. This fact is particularly critical because low-level carbapenem-resistant organisms such as IMP-producing CPE are most prevalent in Japan. We developed a modified selective medium more suitable for IMP-type CPE. METHODS: Fifteen reference CPE strains producing different types of ß-lactamases were used to evaluate the commercially available CHROMagar KPC and chromID CARBA as well as the newly prepared MC-ECC medium (CHROMagar ECC supplemented with meropenem, cloxacillin, and ZnSO4) and M-ECC medium (CHROMagar ECC supplemented with meropenem and ZnSO4). A total of 1035 clinical samples were then examined to detect CPE using chromID CARBA and M-ECC medium. RESULTS: All tested strains producing NDM-, KPC-, and OXA-48-carbapenemases were successfully cultured in the media employed. Although most of the IMP-positive strains did not grow in CHROMagar KPC, chromID CARBA, or MC-ECC, all tested strains grew on M-ECC. When faecal samples were applied to the media, M-ECC medium allowed the best growth of IMP-type CPE with a significantly higher sensitivity (99.3%) than that of chromID CARBA (13.9%). CONCLUSIONS: M-ECC medium was determined as the most favourable selective medium for the detection of IMP-type CPE as well as other types of CPE.

Activation of nuclear receptor CAR by an environmental pollutant perfluorooctanoic acid.

Perfluorocarboxylic acids (PFCAs) including perfluorooctanoic acid (PFOA) are environmental pollutants showing high accumulation, thermochemical stability and hepatocarcinogenicity. Peroxisome proliferator-activated receptor α is suggested to mediate their toxicities, but the precise mechanism remains unclear. Previous reports also imply a possible role of constitutive androstane receptor (CAR), a key transcription factor for the xenobiotic-induced expression of various genes involved in drug metabolism and disposition as well as hepatocarcinogenesis. Therefore, we have investigated whether PFCAs activate CAR. In wild-type but not Car-null mice, mRNA levels of Cyp2b10, a CAR target gene, were increased by PFOA treatment. PFCA treatment induced the nuclear translocation of CAR in mouse livers. Since CAR activators are divided into two types, ligand-type activators and phenobarbital-like indirect activators, we investigated whether PFCAs are CAR ligands or not using the cell-based reporter gene assay that can detect CAR ligands but not indirect activators. As results, neither PFCAs nor phenobarbital increased reporter activities. Interestingly, in mouse hepatocytes, pretreatment with the protein phosphatase inhibitor okadaic acid prevented an increase in Cyp2b10 mRNA levels induced by phenobarbital as reported, but not that by PFOA. Finally, in human hepatocyte-like HepaRG cells, PFOA treatment increased mRNA levels of CYP2B6, a CAR target gene, as did phenobarbital. Taken together, our present results suggest that PFCAs including PFOA are indirect activators of mouse and human CAR and that the mechanism might be different from that for phenobarbital. The results imply a role of CAR in the hepatotoxicity of PFCAs.

Distinct effects of Hoxa2 overexpression in cranial neural crest populations reveal that the mammalian hyomandibular-ceratohyal boundary maps within the styloid process.

Most gnathostomata craniofacial structures derive from pharyngeal arches (PAs), which are colonized by cranial neural crest cells (CNCCs). The anteroposterior and dorsoventral identities of CNCCs are defined by the combinatorial expression of Hox and Dlx genes. The mechanisms associating characteristic Hox/Dlx expression patterns with the topology and morphology of PAs derivatives are only partially known; a better knowledge of these processes might lead to new concepts on the origin of taxon-specific craniofacial morphologies and of certain craniofacial malformations. Here we show that ectopic expression of Hoxa2 in Hox-negative CNCCs results in distinct phenotypes in different CNCC subpopulations. Namely, while ectopic Hoxa2 expression is sufficient for the morphological and molecular transformation of the first PA (PA1) CNCC derivatives into the second PA (PA2)-like structures, this same genetic alteration does not provoke the transformation of derivatives of other CNCC subpopulations, but severely impairs their development. Ectopic Hoxa2 expression results in the transformation of the proximal Meckel's cartilage and of the malleus, two ventral PA1 CNCCs derivatives, into a supernumerary styloid process (SP), a PA2-derived mammalian-specific skeletal structure. These results, together with experiments to inactivate and ectopically activate the Edn1-Dlx5/6 pathway, indicate a dorsoventral PA2 (hyomandibular/ceratohyal) boundary passing through the middle of the SP. The present findings suggest context-dependent function of Hoxa2 in CNCC regional specification and morphogenesis, and provide novel insights into the evolution of taxa-specific patterning of PA-derived structures.

BioFNet: biological functional network database for analysis and synthesis of biological systems.

In synthetic biology and systems biology, a bottom-up approach can be used to construct a complex, modular, hierarchical structure of biological networks. To analyze or design such networks, it is critical to understand the relationship between network structure and function, the mechanism through which biological parts or biomolecules are assembled into building blocks or functional networks. A functional network is defined as a subnetwork of biomolecules that performs a particular function. Understanding the mechanism of building functional networks would help develop a methodology for analyzing the structure of large-scale networks and design a robust biological circuit to perform a target function. We propose a biological functional network database, named BioFNet, which can cover the whole cell at the level of molecular interactions. The BioFNet takes an advantage in implementing the simulation program for the mathematical models of the functional networks, visualizing the simulated results. It presents a sound basis for rational design of biochemical networks and for understanding how functional networks are assembled to create complex high-level functions, which would reveal design principles underlying molecular architectures.

Development of an accurate kinetic model for the central carbon metabolism of Escherichia coli.

BACKGROUND: A kinetic model provides insights into the dynamic response of biological systems and predicts how their complex metabolic and gene regulatory networks generate particular functions. Of many biological systems, Escherichia coli metabolic pathways have been modeled extensively at the enzymatic and genetic levels, but existing models cannot accurately reproduce experimental behaviors in a batch culture, due to the inadequate estimation of a specific cell growth rate and a large number of unmeasured parameters. RESULTS: In this study, we developed a detailed kinetic model for the central carbon metabolism of E. coli in a batch culture, which includes the glycolytic pathway, tricarboxylic acid cycle, pentose phosphate pathway, Entner-Doudoroff pathway, anaplerotic pathway, glyoxylate shunt, oxidative phosphorylation, phosphotransferase system (Pts), non-Pts and metabolic gene regulations by four protein transcription factors: cAMP receptor, catabolite repressor/activator, pyruvate dehydrogenase complex repressor and isocitrate lyase regulator. The kinetic parameters were estimated by a constrained optimization method on a supercomputer. The model estimated a specific growth rate based on reaction kinetics and accurately reproduced the dynamics of wild-type E. coli and multiple genetic mutants in a batch culture. CONCLUSIONS: This model overcame the intrinsic limitations of existing kinetic models in a batch culture, predicted the effects of multilayer regulations (allosteric effectors and gene expression) on central carbon metabolism and proposed rationally designed fast-growing cells based on understandings of molecular processes.

Immunogenicity and Efficacy of A/H1N1pdm Vaccine Among Subjects With Severe Motor and Intellectual Disability in the 2010/11 Influenza Season.

BACKGROUND: While the immunogenicity and effectiveness of seasonal influenza vaccines among subjects with severe motor and intellectual disability (SMID) are known to be diminished, the efficacy of the A/H1N1pdm vaccine has not been evaluated. METHODS: We prospectively evaluated 103 subjects with SMID (mean age, 41.7 years) who received trivalent inactivated influenza vaccine during the 2010/11 influenza season. The hemagglutination inhibition (HI) antibody titer was measured in serum samples collected pre-vaccination (S0), post-vaccination (S1), and end-of-season (S2) to evaluate subjects' immunogenicity capacity. Vaccine efficacy was assessed based on antibody efficacy and achievement proportion. RESULTS: The proportions of seroprotection and seroconversion, and the geometric mean titer (GMT) ratio (GMT at S1/GMT at S0) for A/H1N1pdm were 46.0%, 16.0%, and 1.8, respectively-values which did not meet the European Medicines Evaluation Agency criteria. The achievement proportion was 26%. During follow-up, 11 of 43 subjects with acute respiratory illness were diagnosed with type A influenza according to a rapid influenza diagnostic test (RIDT), and A/H1N1pdm strains were isolated from the throat swabs of 5 of those 11 subjects. When either or both RIDT-diagnosed influenza or serologically diagnosed influenza (HI titer at S2/HI titer at S1 ≥2) were defined as probable influenza, subjects with A/H1N1pdm seroprotection were found to have a lower incidence of probable influenza (odds ratio, 0.31; antibody efficacy, 69%; vaccine efficacy, 18%). CONCLUSIONS: In the present seasonal assessment, antibody efficacy was moderate against A/H1N1pdm among SMID subjects, but vaccine efficacy was low due to the reduced immunogenicity of SMID subjects.

Varicella zoster virus antibody detection: A comparison of four commonly used techniques.

BACKGROUND: Antibody tests for the varicella zoster virus (VZV) include neutralization, fluorescent antibody to membrane antigen (FAMA), immune adherence hemagglutination (IAHA), enzyme immunoassay (EIA), glycoprotein-based enzyme-linked immunosorbent assay (gpELISA), and complement fixation (CF) tests. Of these, FAMA is considered the most sensitive. However, in Japan, the EIA method is most frequently employed. OBJECTIVE: The VZV antibody detection rate of the FAMA, EIA, gpELISA, and IAHA methods was compared. METHODS: Four types of antibody tests were conducted with sera collected from 83 college students. The relationships between two antibody tests were examined using Pearson's correlation coefficients. RESULTS: All 83 subjects were observed to be VZV antibody-positive using the FAMA method. The Pearson correlation coefficients of gpELISA, EIA, and IAHA relative to FAMA were 0.808, 0.782, and 0.356, respectively. The positive agreement rate of IAHA relative to FAMA was 88.0% (73/83), whereas those of gpELISA and EIA were both 97.6% (81/83). Furthermore, EIA showed 100% positive agreement with gpELISA and a high correlation coefficient of 0.911, whereas these values for IAHA compared to gpELISA were much lower (90.1% and 0.530). The calculated Pearson correlation coefficient for comparison of the EIA and IAHA methods was 0.498, with a positive agreement rate of 90.1% (73/81). CONCLUSIONS: The EIA method should be employed in Japan based on the similarity of the positivity between EIA and gpELISA, as it is more available and practical than gpELISA.

Biosynthesis of D-aspartate in mammals: the rat and human homologs of mouse aspartate racemase are not responsible for the biosynthesis of D-aspartate.

D-Aspartate (D-Asp) has important physiological functions, and recent studies have shown that substantial amounts of free D-Asp are present in a wide variety of mammalian tissues and cells. Biosynthesis of D-Asp has been observed in several cultured rat cell lines, and a murine gene (glutamate-oxaloacetate transaminase 1-like 1, Got1l1) that encodes Asp racemase, a synthetic enzyme that produces D-Asp from L-Asp, was proposed recently. The product of this gene is homologous to mammalian glutamate-oxaloacetate transaminase (GOT). Here, we tested the hypothesis that rat and human homologs of mouse GOT1L1 are involved in Asp synthesis. The following two approaches were applied, since the numbers of attempts were unsuccessful to prepare soluble GOT1L1 recombinant proteins. First, the relationship between the D-Asp content and the expression levels of the mRNAs encoding GOT1L1 and D-Asp oxidase, a primary degradative enzyme of D-Asp, was examined in several rat and human cell lines. Second, the effect of knockdown of the Got1l1 gene on D-Asp biosynthesis during culture of the cells was determined. The results presented here suggest that the rat and human homologs of mouse GOT1L1 are not involved in D-Asp biosynthesis. Therefore, D-Asp biosynthetic pathway in mammals is still an urgent issue to be resolved.

Influenza A(H1N1)pdm09 vaccine effectiveness and other characteristics associated with hospitalization in chronic liver disease patients.

BACKGROUND & AIMS: To date, few studies have investigated the clinical effectiveness of influenza vaccine in chronic liver disease patients. The aim of this study was to examine the effectiveness of monovalent inactivated influenza A(H1N1)pdm09 vaccine and other characteristics associated with hospitalization in patients with chronic hepatitis C. METHODS: We conducted a hospital-based cohort study during influenza A(H1N1)pdm09 pandemic. A total of 408 patients (132 vaccinated, 276 unvaccinated) with detectable HCV-RNA were followed up with respect to any hospitalization using a weekly postal questionnaire. Reported hospitalizations were verified by medical records. RESULTS: During the epidemic period, 28 hospitalizations (6 vaccinated, 22 unvaccinated) were observed. After adjustment for potential confounders, vaccination decreased the odds ratio (OR) for hospitalization with marginal significance (OR = 0.43, 95%CI = 0.16-1.17). Besides, positive association with hospitalization was observed in patients with albumin levels <3.5 g/dl (OR = 8.40, 3.66-19.3) and steroid users (OR = 5.58, 0.98-31.7). CONCLUSIONS: Among patients with chronic hepatitis C, A(H1N1)pdm09 vaccine appeared to have a protective effect against hospitalization. Those patients with a higher risk for hospitalization should be carefully followed during the influenza season, even when vaccinated.

CADLIVE toolbox for MATLAB: automatic dynamic modeling of biochemical networks with comprehensive system analysis.

Mathematical modeling has become a standard technique to understand the dynamics of complex biochemical systems. To promote the modeling, we had developed the CADLIVE dynamic simulator that automatically converted a biochemical map into its associated mathematical model, simulated its dynamic behaviors and analyzed its robustness. To enhance the feasibility by CADLIVE and extend its functions, we propose the CADLIVE toolbox available for MATLAB, which implements not only the existing functions of the CADLIVE dynamic simulator, but also the latest tools including global parameter search methods with robustness analysis. The seamless, bottom-up processes consisting of biochemical network construction, automatic construction of its dynamic model, simulation, optimization, and S-system analysis greatly facilitate dynamic modeling, contributing to the research of systems biology and synthetic biology. This application can be freely downloaded from http://www.cadlive.jp/CADLIVE_MATLAB/ together with an instruction.

PredIL13: Stacking a variety of machine and deep learning methods with ESM-2 language model for identifying IL13-inducing peptides.

Interleukin (IL)-13 has emerged as one of the recently identified cytokine. Since IL-13 causes the severity of COVID-19 and alters crucial biological processes, it is urgent to explore novel molecules or peptides capable of including IL-13. Computational prediction has received attention as a complementary method to in-vivo and in-vitro experimental identification of IL-13 inducing peptides, because experimental identification is time-consuming, laborious, and expensive. A few computational tools have been presented, including the IL13Pred and iIL13Pred. To increase prediction capability, we have developed PredIL13, a cutting-edge ensemble learning method with the latest ESM-2 protein language model. This method stacked the probability scores outputted by 168 single-feature machine/deep learning models, and then trained a logistic regression-based meta-classifier with the stacked probability score vectors. The key technology was to implement ESM-2 and to select the optimal single-feature models according to their absolute weight coefficient for logistic regression (AWCLR), an indicator of the importance of each single-feature model. Especially, the sequential deletion of single-feature models based on the iterative AWCLR ranking (SDIWC) method constructed the meta-classifier consisting of the top 16 single-feature models, named PredIL13, while considering the model's accuracy. The PredIL13 greatly outperformed the-state-of-the-art predictors, thus is an invaluable tool for accelerating the detection of IL13-inducing peptide within the human genome.