Search

Occupational lung disease manifests complex radiologic findings which have long been a challenge for computer-assisted diagnosis (CAD). This journey started in the 1970s when texture analysis was developed and applied to diffuse lung disease. Pneumoconiosis appears on radiography as a combination of small opacities, large opacities, and pleural shadows. The International Labor Organization International Classification of Radiograph of Pneumoconioses has been the main tool used to describe pneumoconioses and is an ideal system that can be adapted for CAD using artificial intelligence (AI). AI includes machine learning which utilizes deep learning or an artificial neural network. This in turn includes a convolutional neural network. The tasks of CAD are systematically described as classification, detection, and segmentation of the target lesions. Alex-net, VGG16, and U-Net are among the most common algorithms used in the development of systems for the diagnosis of diffuse lung disease, including occupational lung disease. We describe the long journey in the pursuit of CAD of pneumoconioses including our recent proposal of a new expert system.

This study investigated the relationship between cerebral gray matter (GM) regions and driving safety behaviors (DSBs) of 98 older drivers without dementia (mean age, 77.72 ± 3.677 years). Their DSBs were evaluated on actual vehicles running on a closed-circuit course. The DSB was scored in six categories: DSB1, visual search behavior; DSB2, speeding; DSB3, signaling of the indicator; DSB4, vehicle stability; DSB5, positioning; and DSB6, steering. The scores were calculated by a driving instructor; larger scores indicated safer driving performances. Regional GM volumes were measured with voxel-based morphometry by magnetic resonance imaging (MRI). In 56 GM regions, 18 regions were correlated with DSB categories except for DSB4. When a single GM region was correlated with multiple DSB categories, a positive or negative response was uniformly determined for the respective region despite clear differences in the DSB categories. This result suggests the existence of two contradictory mechanisms in the brain for DSB. The left postcentral gyrus may largely function in regulating DSBs because it was negatively correlated with five of six DSB categories. Thus, MRI’s measurement of regional GM volumes may help deepen the understanding of the diversity and complexity inherent in brain functions for DSBs.

The Keras deep learning framework is employed to study MRI brain data in a preliminary analysis of brain structure using a convolutional neural network. The results obtained are matched with the content of personality questionnaires. The Big Five personality traits provide easy differentiation for dividing personalities into different groups. Until now, the highest accuracy obtained from the results of personality prediction from the analysis of brain structure is about 70%. Although there is still no effective evidence to prove a clear relationship between brain structure and personality, the obtained results could prove helpful in understanding the basic relationship between brain structure and personality characteristics.

The diagnosis of brain pathologies usually involves imaging to analyze the condition of the brain. Magnetic resonance imaging (MRI) technology is widely used in brain disorder diagnosis. The image quality of MRI depends on the magnetostatic field strength and scanning time. Scanners with lower field strengths have the disadvantages of a low resolution and high imaging cost, and scanning takes a long time. The traditional super-resolution reconstruction method based on MRI generally states an optimization problem in terms of prior information. It solves the problem using an iterative approach with a large time cost. Many methods based on deep learning have emerged to replace traditional methods. MRI super-resolution technology based on deep learning can effectively improve MRI resolution through a three-dimensional convolutional neural network, however, the training costs are relatively high. In this paper, we propose the use of two-dimensional super-resolution technology for the super-resolution reconstruction of MRI images. In the first reconstruction, we choose a scale factor of 2 and simulate half the volume of MRI slices as input. We utilize a receiving field block enhanced super-resolution generative adversarial network (RFB-ESRGAN), which is superior to other super-resolution technologies in terms of texture and frequency information. We then rebuild the super-resolution reconstructed slices in the MRI. In the second reconstruction, the image after the first reconstruction is composed of only half of the slices, and there are still missing values. In our previous work, we adopted the traditional interpolation method, and there was still a gap in the visual effect of the reconstructed images. Therefore, we propose a noise-based super-resolution network (nESRGAN). The noise addition to the network can provide additional texture restoration possibilities. We use nESRGAN to further restore MRI resolution and high-frequency information. Finally, we achieve the 3D reconstruction of brain MRI images through two super-resolution reconstructions. Our proposed method is superior to 3D super-resolution technology based on deep learning in terms of perception range and image quality evaluation standards.

Research on medical image recognition by machine learning has been studied for a long time. Medical image recognition has been mainly for detection of diseases or symptoms. In recent years, there is a hypothesis which human health conditions and personal attributes affect brain structure. Most studies focus on T1-weighted images, however T2-weighted and diffusion-weighted images are also used in medical practice. Diffusion images contains signals and information which are possibly used to discriminate human health or other state rather and are different from T1-weighted or T2-weighted images. Some studies using 3D-CNN (3-dimensional convolutional neural network) have been applied for brain structural magnetic resonance imaging such as T1-weighted images, but diffusion tensor images cannot be directly input to 3D-CNN because their data is represented as the tensor of the movement of water molecules. Therefore, in this research, we propose 3D-CNN which input the diffusion images of each direction to different channel. We construct multi-channel 3D-CNN and verify the effect of brain structure image mode on gender discrimination using T1, T2, and DTI.

Vegetable mass estimation from monocular RGB camera images is proposed. Vegetables are fragmented and placed on a conveyor belt of food processing machine and the monocular camera placed over the belt take pictures of vegetables on the belt. The proposed system does not employ any scale, load cell, and other mass scaling equipment. We apply pre-trained convolutional neural networks to estimate the mass of vegetables. Transfer learning including various levels of fine-tuning is also applied. For pre-trained network, we use Xception, VGG16, ResNet50, and Inception_v3, which are pre-trained using ImageNet. The result shows that the best estimation accuracy is achieved by VGG16, whose MAPE (mean average percentage error) is 11.1%. Additionally, we fine-tune VGG16 and the accuracy reduces to 7.9% for MAPE. From this result, the performance of CNN model can improve by fine-tuning. The proposed system can be applied to low-cost, high-speed, and efficient measurement of foods replaced to load cells.

Magnetic resonance imaging (MRI) is one of the most important diagnostic imaging methods, which is widely used in diagnosis and image-guided therapy, especially imaging diagnosis of the brain. However, MRI images have the characteristics of low resolution, and there are limitations such as long imaging time and noise. Super-resolution techniques have been studied on three-dimensional MRI images using three-dimensional convolutional neural network. Based on some related techniques of super-resolution reconstruction of two-dimensional MRI slices, we evaluated the capability of several super-resolution technologies. We utilize the super-resolution algorithm based on generative adversarial network ESRGAN to realize super-resolution reconstruction of two-dimensional MRI slices, and then we further demonstrate that frequent details can be obtained from ESRGAN. In the aspect of two-dimensional to three-dimensional reconstruction, we use the technique of two-dimensional super-resolution on slices from three different latitudes. We rebuild reconstructed two-dimensional images into a three-dimensional form. Then based on the principle of linear interpolation, we use the surrounding effective pixel values to interpolate the null value of each slice, and realize the reconstruction of three-dimensional brain MRI.

Introduction To investigate the biodistribution and retention properties of the new super paramagnetic iron oxide (new SPIO: mean hydrodynamic diameter, 100 nm) nanoparticles, which have concentrated polymer brushes in the outer shell and are difficult for phagocytes to absorb, and to compare the new SPIO with clinically approved SPIO (Resovist: mean hydrodynamic diameter, 57 nm). Materials and methods 16 male C57BL/6N mice were divided in two groups according to the administered SPIO (n = 8 for each group; intravenous injection does, 0.1 ml). In vivo magnetic resonance imaging (MRI) was performed before and one hour, one day, one week and four weeks after SPIO administration by two dimensional-the fast low angle shot (2D-FLASH) sequence at 11.7T. Ex vivo high-resolution images of fixed organs were also obtained by (2D-FLASH). After the ex vivo MRI, organs were sectioned and evaluated histologically to confirm the biodistribution of each particle precisely. Results The new SPIO was taken up in small amounts by liver Kupffer cells and showed a unique in vivo MRI contrast pattern in the kidneys, where the signal intensity decreased substantially in the boundaries between cortex and outer medulla and between outer and inner medulla. We found many round dark spots in the cortex by ex vivo MRI in both groups. Resovist could be detected almost in the cortex. The shapes of the dark spots were similar to those observed in the new SPIO group. Transmission electron microscopy revealed that Resovist and the new SPIO accumulated in different cells of glomeruli, that is, endothelial and mesangial cells, respectively. Conclusion The new SPIO was taken up in small amounts by liver tissue and showed a unique MRI contrast pattern in the kidney. The SPIO were found in the mesangial cells of renal corpuscles. Our results indicate that the new SPIO may be potentially be used as a new contrast agent for evaluation of kidney function as well as immunune function.

We have validated that the electrical performances of the In0.53Ga0.47As MOSFETs such as sub-threshold slope (SS) and electron mobility were dependent on interfacial reactions in the metal/highk/InGaAs gate stacks which could be controlled remotely by choice of the metal electrodes. We demonstrated In0.53Ga0.47As MOSFETs with high mobility (peak mobility ~1300 cm2/Vs) and superior SS performance (SS 76.4 mV/dec) at the scaled CET region owing to the remote reduction of the native III-V oxide by the TiN electrodes.

The use of Sn-3mass%Ag-0.5mass%Cu lead-free solder (SAC305) has become common. Since SAC305 has a higher content of tin than conventional tin&ndash, lead eutectic solder, erosion of the Fe plating layer used in the solder iron tip and the point soldering machine nozzle frequently occurs. In this study, to prolong the life of the Fe plating layer, the applicability of composite plating in which a carbon-type filler is compounded with Fe was studied. Graphite and a multi-walled carbon nanotube (MWCNT) were used as filler materials in the composite plating layer. For both Fe-graphite and Fe-MWCNT composite plating layers, solderability testing and erosion-resistance testing were carried out. In the solderability test, although the spread rates of SAC305 to both Fe-graphite and Fe-MWCNT plating layers slightly decreased compared to the Fe plating layer, SAC305 solder was not repelled against both plating layers. In the erosion-resistance test, the Fe-MWCNT composite plating layer performed the best with the least erosion depth. The erosion depth of the Fe-graphite composite plating layer and the Fe plating layer were 10 and 100 times larger than that of the Fe-MWCNT composite plating layer, respectively. It was confirmed that the diffusion of Fe into molten SAC305 could be greatly reduced due to the composing carbon filler in Fe.

Ascorbate (AsA) is required as a cofactor and is widely distributed in plants and animals. Recently, it has been suggested that the nematode Caenorhabditis elegans also synthesizes AsA. However, its biosynthetic pathway is still unknown. To further understand AsA biosynthesis in C. elegans, we analyzed the incorporation of the 13C atom into AsA using gas chromatography-mass spectrometry (GC-MS) in worms fed with D-Glc (1-13C)-labeled Escherichia coli. GC-MS analysis revealed that AsA biosynthesis in C. elegans, similarly to that in mammalian systems, involves carbon skeleton rearrangement. The addition of L-gulono-1,4-lactone, an AsA precursor in the mammalian pathway, significantly increased AsA level in C. elegans, whereas the addition of L-galactono-1,4-lactone, an AsA precursor in the plant and Euglena pathway, did not affect AsA level. The suppression of E03H4.3 (an ortholog of gluconolactonase) or the deficiency of F54D5.12 (an ortholog of L-gulono-1,4-lactone oxidase) significantly decreased AsA level in C. elegans. Although N2- and AsA-deficient F54D5.12 knockout mutant worm (tm6671) morphologies and the ratio of collagen to non-collagen protein did not show any significant differences, the mutant worms exhibited increased malondialdehyde levels and reduced lifespan compared with the N2 worms. In conclusion, our findings indicate that the AsA biosynthetic pathway is similar in C. elegans and mammals.

The authors have developed fault displacement (FD) probabilistic risk assessment (PRA) methodology, through application to a hypothetical nuclear power plant (NPP). Core damage frequency (CDF) and its uncertainty of a hypothetical NPP for dip-slip faults and strike-slip faults were evaluated using a probabilistic FD hazard, FD fragilities of building and components, and fault tree and event tree models. Important initiating events, accident sequences, components and structure failures, as well as the range of FD that characterize the core damage risk profile were identified. Through the study, the authors confirmed the feasibility of the methodology, and also identified the important source of uncertainties within the methodology that require further development. Risk insights obtained from the FD PRA can be used to investigate countermeasures to reduce the FD risk.

An innovative engineer who can address a social challenge using big data processing, AI and cloud computing technologies with the generation of new business and value is required from industry. enPiT is an education project to develop the advanced IT engineer based on practical education in cooperation between industry and academia promoted by Ministry of Education, Culture, Sports, Science and Technology (MEXT) of Japan. In this paper, we introduce how we designed PBL centered curriculum named AiBiC Spiral under the framework of enPiT education project, and analyze educational effect of our program based on work products and the questionnaire result targeting students who took and completed to the program of 2017.

Graphical abstractDisplay Omitted HighlightsWe propose a cooperative approach for numerical optimization.A fast operator is proposed to capture the interdependencies among variables.Problem decomposition is performed based on the interdependencies.A cross-cluster mutation is proposed to optimize the subproblems.This approach is effective for large scale optimization problems. Cooperative optimization algorithms have been applied with success to solve many optimization problems. However, many of them often lose their effectiveness and advantages when solving large scale and complex problems, e.g., those with interacted variables. A key issue involved in cooperative optimization is the task of problem decomposition. In this paper, a fast search operator is proposed to capture the interdependencies among variables. Problem decomposition is performed based on the obtained interdependencies. Another key issue involved is the optimization of the subproblems. A cross-cluster mutation strategy is proposed to further enhance exploitation and exploration. More specifically, each operator is identified as exploitation-biased or exploration-biased. The population is divided into several clusters. For the individuals within each cluster, the exploitation-biased operators are applied. For the individuals among different clusters, the exploration-biased operators are applied. The proposed operators are incorporated into the original differential evolution algorithm. The experiments were carried out on CEC2008, CEC2010, and CEC2013 benchmarks. For comparison, six algorithms that yield top ranked results in CEC competition are selected. The comparison results demonstrated that the proposed algorithm is robust and comprehensive for large scale optimization problems.

This paper proposes two deep learning modes which combine sparse auto-encoder with extreme learning machine (ELM) and kernel extreme learning machine (KELM), namely as Stacked Sparse Auto-encoder-Extreme Learning Machine and Stacked Sparse Auto-encoder-Kernel Extreme Learning Machine. The proposed models are applied to the image recognition task. To learn features from the original input data, SSAE with deep architecture is employed. Then, to construct a unified neural network learning model, ELM and KELM are selected to classify the extracted features. To evaluate the performance of the two proposed models, we carry out experiments on three different image data sets respectively. The results show that the proposed models’ performance can be not only superior to the shallow architecture models of Support Vector Machine, ELM and KELM but also better than the deep architecture models, such as Stacked Auto-encoder, Deep Belief Network and Stacked Denoising Auto-encoder.

Continual progress in the fields of computer vision and machine learning has provided opportunities to develop automatic tools for tagging images; this facilitates searching and retrieving. However, due to the complexity of real-world image systems, effective and efficient image annotation is still a challenging problem. In this paper, we present an annotation technique based on the use of image content and word correlations. Clusters of images with manually tagged words are used as training instances. Images within each cluster are modeled using a kernel method, in which the image vectors are mapped to a higher-dimensional space and the vectors identified as support vectors are used to describe the cluster. To measure the extent of the association between an image and a model described by support vectors, the distance from the image to the model is computed. A closer distance indicates a stronger association. Moreover, word-to-word correlations are also considered in the annotation framework. To tag an image, the system predicts the annotation words by using the distances from the image to the models and the word-to-word correlations in a unified probabilistic framework. Simulated experiments were conducted on three benchmark image data sets. The results demonstrate the performance of the proposed technique, and compare it to the performance of other recently reported techniques. HighlightsWe propose a support vector based approach to tag images.In the training process, training images are modeled using a kernel method.The training models can delineate arbitrary organized training images.In the tagging process, tags are predicted using a probabilistic approach.This approach is effective when the training images are irregularly organized.

Fucoidan and chondroitin sulfate, which are well known sulfated polysaccharides, were depolymerized under hydrothermal conditions (120-180°C, 5-60min) as a method for the preparation of sulfated polysaccharides with controlled molecular weights. Fucoidan was easily depolymerized, and the change of the molecular weight values depended on the reaction temperature and time. The degree of sulfation and IR spectra of the depolymerized fucoidan did not change compared with those of untreated fucoidan at reaction temperatures below 140°C. However, fucoidan was partially degraded during depolymerization above 160°C. Nearly the same depolymerization was observed for chondroitin sulfate. These results indicate that hydrothermal treatment is applicable for the depolymerization of sulfated polysaccharides, and that low molecular weight products without desulfation and deformation of the initial glycan structures can be obtained under mild hydrothermal conditions.

Auditory hallucinations and delusions are core symptoms of schizophrenia, which interact with each other. The attribution of auditory hallucinations to other people is considered to lead to secondary delusions. This study examined whether brief psychoeducation can change the cognition of auditory hallucinations, particularly, their attribution, and thus alleviate secondary delusions. Twenty-two schizophrenic patients with auditory hallucinations were recruited in this open study. The intervention consisted of five sessions during the course of 4 weeks. Outcome measures were used to assess delusions, beliefs about auditory hallucinations, and depression. At the end of the intervention, statistically significant reduction was observed in both delusions and depression. Beliefs about hallucinations showed statistically significant improvement in terms of malevolence, omnipotence, and resistance but not in terms of benevolence and engagement. In conclusion, the present study suggests that psychoeducation might be useful in reducing secondary delusions without exacerbating a depressive state.

Sensing bacteria such as in the medical front without their cultivation is attracting a great deal of interest to avoid an outbreak of bacterial infection because definitive diagnosis takes a long time [1-3]. For the quick test in medical front, we have been developing a portable small sensor that can detect pathogenic bacteria, Legionella, by detecting their fluorescence on a chip without their cultivation [4, 5]. There are about 50 species in the genus Legionella [6]. Some of them are known to emit fluorescence when they are irradiated with ultraviolet (UV) light at around 360 nm of wavelength. Applying an optical sensor therefore can detect and identify each species of Legionella. This work describes our recent researches on bacterial identification using a photogate-type optical sensor. Our used bacteria in this study are Legionella dumoffii (L. dumoffii) and Legionella erythra (L. erythra) which respectively emit blue and red fluorescence when they are irradiated by UV-light as shown in Fig. 1. The schematic cross section of photogate-type optical sensor fabricated in our lab. is illustrated in Fig. 2. Fluorescence emitted from Legionella entered to a depth W in the sensing area through the photogate generates photocurrent, Iout according to the following formula, Iout = qSλ/hc x (1-exp(-αW) ) x Φ0(t), q: elementary charge, S: sensing area, λ: wavelength of fluorescence, h: Planck constant, C: light velocity, α: absorption coefficient of light with wavelength of λ in Si, W: penetration depth of light with wavelength of λ, and Φ0(t): intensity of the fluorescence at time, t [7-9]. Obtaining the characteristic fluorescent wavelength, λ and the UV-photoirradiation time dependence of its photointensity, Φ0(t) enables to identify the species of Legionella. The characteristic wavelengths were selected from the fluorescent spectra obtained using a combination of fluorescence microscope and spectrometer between 400 nm and 800 nm of wavelengths both for L. dumoffii and L. erythra. The UV-photoirradiation time dependence of the fluorescent intensities at characteristic wavelengths was then obtained both for L. dumoffii and L. erythra. For the L. dumoffii case, the fluorescent intensity at its characteristic wavelength of 470 nm rapidly decreased with the photoirradiation time, meaning a fluorescent material in the cells of L. dumoffii changes to a non-fluorescent material by UV-photoreaction. For the L. erythra case, on the other hand, the fluorescent intensity at its characteristic wavelength of 675 nm gradually increased to saturate with the photoirradiation time, meaning the production of fluorescent material began to occur when L. erythra was irradiated by the UV-light. Thus, the fluorescent behavior of Legionella under UV-photoirradiation differs largely each other. Finally, the photocurrent generated by the fluorescence emitted from L. dumoffii and L. erythra at their characteristic wavelengths were measured by using the photogate-type optical sensor schematically shown in Fig. 2. The UV-photoirradiation time dependences of the photocurrent were found to be in good agreement with those of fluorescent intensity obtained by the combination of fluorescence microscope and spectrometer, showing that the photogate-type optical sensor is able to be used for bacterial identification instead of a bulky system of spectrometer and fluorescence microscope. This will pave the way for portable bacterial sensors for use in medical front etc. References T. Sunagawa, T. Saito, K. Kinoshita, K. Nakajima, and K. Oishi, ISAR, 34, 160 (2013). T. Karasutani, K. Arai, J. Isobe, J. Kanaya, K. Ogata, S. Izumiyama, K. Yagita, T. Yazaki, M. Yoshizaki, and F. Kura, ISAR, 34, 165 (2013). B. S. Fields, R. F. Benson, and R. E. Besser, Clin. Microbiol. Rev., 15, 506 (2002). R. Hayashi, H. Nakazawa, K. Sawada, M. Ishida, H. Ishii, K. Machida, C. Wang, K. Iida, M. Saito, and S. Yoshida, Ext. Abstract No. 1485, The 225th Electrochemical Society Spring Meeting, Orlando, USA (2014). Y. Nishimura, M. Ishida, K. Sawada, H. Ishii, K. Machida, K. Masu, C. Wang, K. Iida, M. Saito, and S. Yoshida, ECS Trans., 69, 259 (2015). J. A. Maekawa, Y. Hayakawa, H. Sugie, A. Moribayashi, F. Kura, B. Chang, A. Wada, and H. Watanabe, J. Biochem. Biophys. Res. Commun., 323, 954 (2004). Y. Maruyama, M. Ishida, and K. Sawada, Jpn. J. Appl. Phys., 48, 067003 (2009). Y. Moriwaki, K. Takahashi, I. Akita, M. Ishida, and K. Sawada, Jpn .J. Appl. Phys., 54, 04DL03 (2015). Y. J. Choi, K. Takahashi, M. Matsuda, T. Hizawa, Y. Moriwaki, F. Dasai, Y. Kimura, I. Akita, T. Iwata, M. Ishida, and K. Sawada, Jpn. J. Appl. Phys., 55, 04EM10 (2016). Figure 1

Formation of III–V oxides at high-k/InGaAs interface was systematically investigated by means of physical and electrical characterizations. We found that, despite previous reports, a well-controlled remote oxidation of InGaAs surfaces via metal/high-k stacks effectively passivates electrical defects at high-k/InGaAs interfaces. Ultrathin regrown III–V oxides (~0.5 nm) caused by oxygen diffusion from the thin Pd (10 nm thick) electrode proved beneficial in achieving superior interface quality through detailed studies with varying structures of metal electrodes. The chemical bonding states of the ultrathin regrown III–V oxides were identified as In–As–O and Ga–As–O bonds, and there were rarely direct bonds between indium or gallium and oxygen at the initial oxidation reaction. On the basis of these findings, we propose a remote oxidation model at the high-k/InGaAs interfaces and discuss physical origins of the improved and degraded interface quality caused by the interlayer growth.

Protein A affinity chromatography is the standard purification process for the capture of therapeutic antibodies. The individual IgG-binding domains of protein A (E, D, A, B, C) have highly homologous amino acid sequences. From a previous report, it has been assumed that the C domain has superior resistance to alkaline conditions compared to the other domains. We investigated several properties of the C domain as an IgG-Fc capture ligand. Based on cleavage site analysis of a recombinant protein A using a protein sequencer, the C domain was found to be the only domain to have neither of the potential alkaline cleavage sites. Circular dichroism (CD) analysis also indicated that the C domain has good physicochemical stability. Additionally, we evaluated the amino acid substitutions at the Gly-29 position of the C domain, as the Z domain (an artificial B domain) acquired alkaline resistance through a G29A mutation. The G29A mutation proved to increase the alkaline resistance of the C domain, based on BIACORE analysis, although the improvement was significantly smaller than that observed for the B domain. Interestingly, a number of other amino acid mutations at the same position increased alkaline resistance more than did the G29A mutation. This result supports the notion that even a single mutation on the originally alkali-stable C domain would improve its alkaline stability. An engineered protein A based on this C domain is expected to show remarkable performance as an affinity ligand for immunoglobulin.

Aims To improve the efficiency of asymmetric hydrolysis of 3-(4-chlorophenyl) glutaric acid diamide (CGD) using a recombinant Comamonas sp. KNK3-7 amidase (CoAM) produced in Escherichia coli. Methods and Results The CoAM gene was cloned, sequenced, and found to comprise 1512 bp, encoding a polypeptide of 54,054 Da. CoAM-transformed E. coli were able to perform R-selective hydrolysis of CGD; however, complete conversion of 166.2 mmol l−1 CGD in 28 h could not be obtained. We attempted to optimize the reactivity of CoAM by mutating single amino acids in the substrate-binding domain. Notably, the methionine-substituted L146M mutant enzyme showed increased reactivity, completing the conversion of 166.2 mmol l−1 CGD in just 4 h. The Km value for L146M was lower than that of CoAM. Conclusions We succeeded in creating the L146M mutant of CoAM with increased substrate affinity and found that this was the best mutant for the hydrolysis of CGD. Significance and Impact of the Study Increasing the efficiency of hydrolysis of 3-substituted glutaric acid diamides is useful to improve the synthesis of optically active 3-substituted gamma-aminobutyric acid, used in the production of pharmacological drugs. This is the first report of efficient hydrolysis of CGD using amidase mutant-producing E. coli cells. This article is protected by copyright. All rights reserved.