Topological magnetic structure generation using VAE-GAN hybrid model and discriminator-driven latent sampling.

Recently, deep generative models using machine intelligence are widely utilized to investigate scientific systems by generating scientific data. In this study, we experiment with a hybrid model of a variational autoencoder (VAE) and a generative adversarial network (GAN) to generate a variety of plausible two-dimensional magnetic topological structure data. Due to the topological properties in the system, numerous and diverse metastable magnetic structures exist, and energy and topological barriers separate them. Thus, generating a variety of plausible spin structures avoiding those barrier states is a challenging problem. The VAE-GAN hybrid model can present an effective approach to this problem because it brings the advantages of both VAE's diversity and GAN's fidelity. It allows one to perform various applications including searching a desired sample from a variety of valid samples. Additionally, we perform a discriminator-driven latent sampling (DDLS) using our hybrid model to improve the quality of generated samples. We confirm that DDLS generates various plausible data with large coverage, following the topological rules of the target system.

Super-resolution of magnetic systems using deep learning.

We construct a deep neural network to enhance the resolution of spin structure images formed by spontaneous symmetry breaking in the magnetic systems. Through the deep neural network, an image is expanded to a super-resolution image and reduced to the original image size to be fitted with the input feed image. The network does not require ground truth images in the training process. Therefore, it can be applied when low-resolution images are provided as training datasets, while high-resolution images are not obtainable due to the intrinsic limitation of microscope techniques. To show the usefulness of the network, we train the network with two types of simulated magnetic structure images; one is from self-organized maze patterns made of chiral magnetic structures, and the other is from magnetic domains separated by walls that are topological defects of the system. The network successfully generates high-resolution images highly correlated with the exact solutions in both cases. To investigate the effectiveness and the differences between datasets, we study the network's noise tolerance and compare the networks' reliabilities. The network is applied with experimental data obtained by magneto-optical Kerr effect microscopy and spin-polarized low-energy electron microscopy.

Searching for the ground state of complex spin-ice systems using deep learning techniques.

Searching for the ground state of a given system is one of the most fundamental and classical questions in scientific research fields. However, when the system is complex and large, it often becomes an intractable problem; there is essentially no possibility of finding a global energy minimum state with reasonable computational resources. Recently, a novel method based on deep learning techniques was devised as an innovative optimization method to estimate the ground state. We apply this method to one of the most complicated spin-ice systems, aperiodic Penrose P3 patterns. From the results, we discover new configurations of topologically induced emergent frustrated spins, different from those previously known. Additionally, a candidate of the ground state for a still unexplored type of Penrose P3 spin-ice system is first proposed through this study. We anticipate that the capabilities of the deep learning techniques will not only improve our understanding on the physical properties of artificial spin-ice systems, but also bring about significant advances in a wide range of scientific research fields requiring computational approaches for optimization.

Optimization of physical quantities in the autoencoder latent space.

We propose a strategy for optimizing physical quantities based on exploring in the latent space of a variational autoencoder (VAE). We train a VAE model using various spin configurations formed on a two-dimensional chiral magnetic system. Three optimization algorithms are used to explore the latent space of the trained VAE. The first algorithm, the single-code modification algorithm, is designed for improving the local energetic stability of spin configurations to generate physically plausible spin states. The other two algorithms, the genetic algorithm and the stochastic algorithm, aim to optimize the global physical quantities, such as topological index, magnetization, energy, and directional correlation. The advantage of our method is that various optimization algorithms can be applied in the latent space containing the abstracted representation constructed by the trained VAE model. Our method based on latent space exploration is utilized for efficient physical quantity optimization.

Estimating the effective fields of spin configurations using a deep learning technique.

The properties of complicated magnetic domain structures induced by various spin-spin interactions in magnetic systems have been extensively investigated in recent years. To understand the statistical and dynamic properties of complex magnetic structures, it is crucial to obtain information on the effective field distribution over the structure, which is not directly provided by magnetization. In this study, we use a deep learning technique to estimate the effective fields of spin configurations. We construct a deep neural network and train it with spin configuration datasets generated by Monte Carlo simulation. We show that the trained network can successfully estimate the magnetic effective field even though we do not offer explicit Hamiltonian parameter values. The estimated effective field information is highly applicable; it is utilized to reduce noise, correct defects in the magnetization data, generate spin configurations, estimate external field responses, and interpret experimental images.

Magnetic Hamiltonian parameter estimation using deep learning techniques.

Understanding spin textures in magnetic systems is extremely important to the spintronics and it is vital to extrapolate the magnetic Hamiltonian parameters through the experimentally determined spin. It can provide a better complementary link between theories and experimental results. We demonstrate deep learning can quantify the magnetic Hamiltonian from magnetic domain images. To train the deep neural network, we generated domain configurations with Monte Carlo method. The errors from the estimations was analyzed with statistical methods and confirmed the network was successfully trained to relate the Hamiltonian parameters with magnetic structure characteristics. The network was applied to estimate experimentally observed domain images. The results are consistent with the reported results, which verifies the effectiveness of our methods. On the basis of our study, we anticipate that the deep learning techniques make a bridge to connect the experimental and theoretical approaches not only in magnetism but also throughout any scientific research.

An innovative magnetic state generator using machine learning techniques.

We propose a new efficient algorithm to simulate magnetic structures numerically. It contains a generative model using a complex-valued neural network to generate k-space information. The output information is hermitized and transformed into real-space spin configurations through an inverse fast Fourier transform. The Adam version of stochastic gradient descent is used to minimize the magnetic energy, which is the cost of our algorithm. The algorithm provides the proper ground spin configurations with outstanding performance. In model cases, the algorithm was successfully applied to solve the spin configurations of magnetic chiral structures. The results also showed that a magnetic long-range order could be obtained regardless of the total simulation system size.

Association of preoperative anemia and perioperative allogenic red blood cell transfusion with oncologic outcomes in patients with nonmetastatic colorectal cancer.

Background: We investigated whether preoperative anemia and perioperative blood transfusion (pbt) are associated with overall survival and recurrence-free survival in patients with nonmetastatic colorectal cancer. Methods: From 1 January 2009 to 31 December 2014, 1003 patients with primary colorectal cancer were enrolled in the study. Perioperative clinical and oncologic outcomes were analyzed based on the presence of preoperative anemia and pbt. Results: Preoperative anemia was found in 468 patients (46.7%). In the anemia and no-anemia groups, pbt was performed in 44% and 15% of patients respectively. Independent predictors for pbt were preoperative anemia, higher American Society of Anesthesiologists score, laparotomy, lengthy operative time, advanced TNM stage, T4 stage, and 30-day morbidity. The use of pbt, but not preoperative anemia, was found to be an independent adverse prognostic factor for overall survival. In terms of recurrence-free survival, the presence of preoperative anemia was similarly not a significant prognostic factor, but the use of pbt was an independent factor for an unfavourable prognosis. Conclusions: The use of pbt, but not preoperative anemia, was independently associated with worse overall and recurrence-free survival in nonmetastatic colorectal cancer. For better oncologic outcomes, our findings indicate a need to reduce the use of blood transfusion during the perioperative period.

The spin structures of interlayer coupled magnetic films with opposite chirality.

Using Monte-Carlo simulations and micromagnetic simulations, we reveal how the spin structural correlation and the skyrmion dynamics are affected by the interlayer coupling in a chiral magnetic bilayer system, in which the two layers have opposite chirality. The interaction through interlayer coupling between chiral magnetic structures influences the static and dynamics properties profoundly. The competition between the Dzyaloshinskii-Moriya interaction and the interlayer interaction allows multiple magnetic structures to be energetically stable, which includes sole skyrmion states (skyrmion appears in only one of the layers) and skyrmion pair states (coupled skyrmions in top and bottom layers). When current driven spin transfer torques are applied to each state, the sole skyrmion state is mainly propelled by a spin transfer torque causing the skyrmion hall effect, but the skyrmion pair state is propelled by a torque from skyrmion-skyrmion interaction and not influenced by the skyrmion hall effect. Also upon application of an external magnetic field, we found the skyrmions in a skyrmion pair state extinguish in an exclusive way, as the annihilation of a skyrmion in one of the layers stabilizes the once paired skyrmion in the other layer, i.e. the skyrmion lattice sites have only one skyrmion in either layer.

Role of HSPA1L as a cellular prion protein stabilizer in tumor progression via HIF-1α/GP78 axis.

The cellular prion protein (PrPC) is associated with metastasis, tumor progression and recurrence; however, the precise mechanisms underlying its action is not well understood. Our study found that PrPC degradation decreased tumor progression in colorectal cancer (CRC). In a CRC cell line and human CRC tissue exposed to hypoxia, induced heat-shock 70-kDa protein-1-like (HSPA1L) expression stabilized hypoxia-inducible factor-1α (HIF-1α) protein and promoted PrPC accumulation and tumorigenicity in vivo. PrPC was degraded via the proteasome pathway mediated by the ubiquitin-protein E3 ligase glycoprotein 78 (GP78), which interacts directly with PrPC. However, hypoxia-induced HSPA1L interacted with GP78 and inhibited its functions. HSPA1L knockdown facilitated the interaction of GP78 and PrPC, thereby increasing PrPC ubiquitination. Thus, GP78 was identified as the ubiquitinase for PrPC, thereby revealing an essential mechanism that controls PrPC levels in CRC. Our results suggest that the HSPA1L/HIF-1α/GP78 axis has a crucial role in PrPC accumulation during tumor progression.

Impact of vitamin D, bisphosphonate, and combination therapy on bone mineral density in kidney transplant patients.

BACKGROUND: Osteoporosis can develop and become aggravated in kidney transplant patients; however, the best preventive options for post-transplantation osteoporosis remain controversial. METHODS: We retrospectively analyzed cohort of 182 renal transplant recipients of mean age 46.7 ± 12.1 years including 47.3% women. Seventy-three patients received neither vitamin D nor bisphosphonate after transplantation (group 1). The other patients were classified into the following 3 groups: calcium plus vitamin D (group 2; n = 40); bisphosphonate (group 3; n = 18); and both regimens (group 4; n = 51). Bone mineral density (BMD) was evaluated by dual-energy X-ray absorptiometry at baseline and at 1 year after transplantation. RESULTS: At 1 year after transplantation, T-scores of the femoral neck and entire femur were significantly decreased in group 1 (-0.23 ± 0.65 [P = .004] and -0.21 ± 0.74 [P = .018], respectively), whereas the lumbar spine was significantly increased in group 4 (0.27 ± 0.79; P = .020). Post hoc analysis demonstrated that the delta T-score was significantly lower in group 1 than in group 4 (P = .009, 0.035, and 0.031 for lumbar spine, femoral neck, and entire femur, respectively). In a multivariate analysis adjusted by age, sex, body mass index, dialysis duration, diabetes, calcineurin inhibitors, estimated glomerular filtration rate, and persistent hyperparathyroidism, both group 2 and group 4 showed protective effects on BMD reduction (odds ratio [OR], 0.165; 95% confidence interval [CI] 0.032-0.845 [P = .031]; and OR, 0.169; 95% CI, 0.045-0.626 [P = .008]; respectively). However, group 3 did not show a protective effect (OR, 0.777; 95% CI, 0.198-3.054; P = .718), because their incidence of persistent hyperparathyroidism after transplantation was significantly higher (50.0%) than the other groups (P < .001). The incidence of bone fractures did not differ among the groups. CONCLUSIONS: Combination therapy with vitamin D and bisphosphonate was the most effective regimen to improve BMD among kidney recipients.

Novel chiral magnetic domain wall structure in Fe/Ni/Cu(001) films.

Using spin-polarized low energy electron microscopy, we discovered a new type of domain wall structure in perpendicularly magnetized Fe/Ni bilayers grown epitaxially on Cu(100). Specifically, we observed unexpected Néel-type walls with fixed chirality in the magnetic stripe phase. Furthermore, we find that the chirality of the domain walls is determined by the film growth order with the chirality being right handed in Fe/Ni bilayers and left handed in Ni/Fe bilayers, suggesting that the underlying mechanism is the Dzyaloshinskii-Moriya interaction at the film interfaces. Our observations may open a new route to control chiral spin structures using interfacial engineering in transition metal heterostructures.

Outcomes of management for potential deceased donors.

BACKGROUNDS: Potential deceased donor management optimization is important for organ recovery maximization. Before optimization, the current state of donor management and predictors for organ recovery require analysis. METHODS: We retrospectively analyzed organ procurement activity and medical management for 2005 to 2010 potential brain death donors at Seoul National University Hospital. RESULTS: Of 316 contacts for potential brain-dead donors, 129 (39.7%) patients were transferred to the donor management team. Among the causes of transfer failure, issues related to proper donor management affected 33%. Expanded criteria donors were 17.9% of transferred donors. Organ recovery was successful in 111 (90.2%) donors. A total of 360 organs were recovered, corresponding to a mean of 2.92 ± 1.37 organs per donor. The absence of organ demand was an important cause of recovery failure among less transplanted organs. Brain death-related complications were identified as follows: acute kidney injury (AKI), defined by AKI network criteria, occurred in 19 (15.4%); cardiopulmonary resuscitation in 5 (3.1%); bacteremia in 12 (9.7%); thrombocytopenia in 24 (19.5%); and diabetes insipidus in 42 (34.1%). AKI was a significant independent risk factor for organ recovery failure in both the liver and kidney (odds ratio [OR] 0.147, 95% confidence interval [0.045, 0.473], P = .001; OR 0.096, 95% confidence interval [0.023, 0.392], P = .001, for kidney and liver, respectively). CONCLUSIONS: Both the transfer success rate and rate of organs transplanted per donor of potential deceased donors remained low in Korea. AKI during potential donor management was a risk factor for kidney and liver recovery failure.

Kidney transplantation using expanded criteria donors in Korea.

BACKGROUND: Use of expanded criteria donor (ECD) grafts seeks to solve the organ shortage. We investigated the current status of donor selection and transplantation outcomes. METHODS: We retrospectively analyzed 791 kidney transplantations performed between 1997 and 2009. An expanded criteria deceased donor (ECDD) was defined as an individual who fulfilled the United Network for Organ Sharing criteria or, the Nyberg criteria. An expanded criteria living donor (ECLD) was determined by fulfillment of 1 or more of 5 criteria. RESULTS: Deceased and living donor kidney transplantations were performed in 228 (28.8%) and 563 (71.2%) cases, respectively. Forty-three cases (18.9%) belonged to the ECDDs. The ECDD group showed a lower posttransplantation 1-year estimated glomerular filtration rate (eGFR) than that of the standard criteria deceased donor (SCDD) group (70.7 ± 19.2 vs 48.6 ± 11.5; P < .001). The ECDDs were allocated to older recipients or recipients with more HLA mismatches than SCDDs. The number of ECLD cases was 173 (30.7%). The proportions of each medical abnormality of living donors were as follows: age older than 60 years (0.5%), hypertension (2.5%), obesity (2.1%), low eGFR (25.9%), proteinuria (0%), and microscopic hematuria (1.4%). The ECLD group showed a lower posttransplantation 1-year eGFR than that of the standard criteria living donor (SCLD) group (66.9 ± 16.0 vs 58.3 ± 11.2; P < .001). Graft survival was not different among the donor types (P = .518). CONCLUSIONS: eCDs were 27.3% of the total kidney donors. Posttransplantation 1-year eGFR was lower in the ECD group. However, there was no difference in the graft survival among the different donor types.

Mitochondrial DNA mutations in a patient with sex reversal and clinical features consistent with Fraser syndrome.

We describe a 20-year-old 46,XY woman, with clinical findings of Fraser syndrome and three mitochondrial DNA (mtDNA) mutations of Leber hereditary optic neuropathy. The patient had microphthalmia, blindness, widely spaced nipples, bifid ureter, syndactyly of the toes, and mental retardation. Sonography showed the presence of a uterus and intra-abdominal gonads. The proband was screened for mtDNA mutations because of chronic gastrointestinal pseudo-obstruction, urinary tract dysmotility, seizures, mental retardation and persistent macrocytosis, as well as the intermittent elevation of methylmalonic acid. Analysis of point mutations by multiplex polymerase chain reaction and allele-specific oligonucleotide dot-blot hybridization revealed three homoplasmic mtDNA mutations, T14484C, T4216C, and T3394C. This represents a unique case with sex reversal, Fraser-like syndrome, and mitochondrial disease.

Oxidative modification of human ceruloplasmin by peroxyl radicals.

Ceruloplasmin (CP), the blue oxidase present in all vertebrates, is the major copper-containing protein of plasma. We investigated oxidative modification of human CP by peroxyl radicals generated in a solution containing 2,2'-azobis(2-amidinopropane) dihydrochloride (AAPH). When CP was incubated with AAPH, the aggregation of proteins was increased in a time- and dose-dependent manner. Incubation of CP with AAPH resulted in a loss of ferroxidase activity. Superoxide dismutase and catalase did not protect the aggregation of CP, whereas hydroxyl radical scavengers such as ethanol and mannitol protected the protein aggregation. The aggregation of proteins was significantly inhibited by the copper chelators, diethyldithiocarbamate and penicillamine. Exposure of CP to AAPH led to the release of copper ions from the enzyme and the generation of protein carbonyl derivatives. Subsequently, when the amino acid composition of CP reacted with AAPH was analyzed, cysteine, tryptophan, methionine, histidine, tyrosine, and lysine residues were particularly sensitive.

Transduction of human catalase mediated by an HIV-1 TAT protein basic domain and arginine-rich peptides into mammalian cells.

Antioxidant enzymes such as superoxide dismutase (SOD) and catalase (CAT) have been considered to have a beneficial effect against various diseases mediated by reactive oxygen species (ROS). Although a variety of modified recombinant antioxidant enzymes have been generated to protect against the oxidative stresses, the lack of their transduction ability into cells resulted in limited ability to detoxify intracellular ROS. To render the catalase enzyme capable of detoxifying intracellular ROS when added extracellularly, cell-permeable recombinant catalase proteins were generated. A human liver catalase gene was cloned and fused with a gene fragment encoding the HIV-1 Tat protein transduction domain (RKKRRQRRR) and arginine-rich peptides (RRRRRRRRR) in a bacterial expression vector to produce genetic in-frame Tat-CAT and 9Arg-CAT fusion proteins, respectively. The expressed and purified fusion proteins can be transduced into mammalian cells (HeLa and PC12 cells) in a time- and dose-dependent manner when added exogenously in culture medium, and transduced fusion proteins were enzymatically active and stable for 60 h. When exposed to H(2)O(2), the viability of HeLa cells transduced with Tat-CAT or 9Arg-CAT fusion proteins was significantly increased. In combination with transduced SOD, transduced catalase also resulted in a cooperative increase in cell viability when the cells were treated with paraquat, an intracellular antioxide anion generator. We then evaluated the ability of the catalase fusion proteins to transduce into animal skin. This analysis showed that Tat-CAT and 9Arg-CAT fusion proteins efficiently penetrated the epidermis as well as the dermis of the subcutaneous layer when sprayed on animal skin, as judged by immunohistochemistry and specific enzyme activities. These results suggest that Tat-CAT and 9Arg-CAT fusion proteins can be used in protein therapy for various disorders related to this antioxidant enzyme.

The changes in the expressions of gamma-aminobutyric acid transporters in the gerbil hippocampal complex following spontaneous seizure.

To identify the roles of gamma-aminobutyric acid (GABA) transporter in epileptogenesis and the recovery mechanisms in spontaneous seizure, a chronological and comparative analysis of GABA transporters (GAT) expression was conducted. GAT-1 immunoreactivity was more strongly detected in the pre-seizure group of seizure sensitive (SS) gerbils than that seen in the seizure resistant group. 30 min postictal, the density of GAT-1 immunoreactivity was significantly decreased in the hippocampal complex, as compared to pre-seizure group. 12 h after seizure on-set, the GAT immunodensity recovered to the pre-seizure level. Following the onset of seizure, GAT-3 immunoreactivity remained unchanged. These results suggest that the increase of GAT-1 expression in the SS gerbil hippocampus may affect epileptogenesis in this animal, and the alteration of immunoreactivity following seizure may be compensatory responses to modulate seizure activity.