A Deep Reinforcement Learning Approach to Droplet Routing for Erroneous Digital Microfluidic Biochips.

Digital microfluidic biochips (DMFBs), which are used in various fields like DNA analysis, clinical diagnosis, and PCR testing, have made biochemical experiments more compact, efficient, and user-friendly than the previous methods. However, their reliability is often compromised by their inability to adapt to all kinds of errors. Errors in biochips can be categorized into two types: known errors, and unknown errors. Known errors are detectable before the start of the routing process using sensors or cameras. Unknown errors, in contrast, only become apparent during the routing process and remain undetected by sensors or cameras, which can unexpectedly stop the routing process and diminish the reliability of biochips. This paper introduces a deep reinforcement learning-based routing algorithm, designed to manage not only known errors but also unknown errors. Our experiments demonstrated that our algorithm outperformed the previous ones in terms of the success rate of the routing, in the scenarios including both known errors and unknown errors. Additionally, our algorithm contributed to detecting unknown errors during the routing process, identifying the most efficient routing path with a high probability.

Pix2Pix-Based Monocular Depth Estimation for Drones with Optical Flow on AirSim.

In this work, we propose a method for estimating depth for an image of a monocular camera in order to avoid a collision for the autonomous flight of a drone. The highest flight speed of a drone is generally approximate 22.2 m/s, and long-distant depth information is crucial for autonomous flights since if the long-distance information is not available, the drone flying at high speeds is prone to collisions. However, long-range, measurable depth cameras are too heavy to be equipped on a drone. This work applies Pix2Pix, which is a kind of Conditional Generative Adversarial Nets (CGAN). Pix2Pix generates depth images from a monocular camera. Additionally, this work applies optical flow to enhance the accuracy of depth estimation. In this work, we propose a highly accurate depth estimation method that effectively embeds an optical flow map into a monocular image. The models are trained with taking advantage of AirSim, which is one of the flight simulators. AirSim can take both monocular and depth images over a hundred meter in the virtual environment, and our model generates a depth image that provides the long-distance information than images captured by a common depth camera. We evaluate accuracy and error of our proposed method using test images in AirSim. In addition, the proposed method is utilized for flight simulation to evaluate the effectiveness to collision avoidance. As a result, our proposed method is higher accuracy and lower error than a state of work. Moreover, our proposed method is lower collision than a state of work.

Robust Self-Adaptation Fall-Detection System Based on Camera Height.

Vision-based fall-detection methods have been previously studied but many have limitations in terms of practicality. Due to differences in rooms, users do not set the camera or sensors at the same height. However, few studies have taken this into consideration. Moreover, some fall-detection methods are lacking in terms of practicality because only standing, sitting and falling are taken into account. Hence, this study constructs a data set consisting of various daily activities and fall events and studies the effect of camera/sensor height on fall-detection accuracy. Each activity in the data set is carried out by eight participants in eight directions and taken with the depth camera at five different heights. Many related studies heavily depended on human segmentation by using Kinect SDK but this is not reliable enough. To address this issue, this study proposes Enhanced Tracking and Denoising Alex-Net (ETDA-Net) to improve tracking and denoising performance and classify fall and non-fall events. Experimental results indicate that fall-detection accuracy is affected by camera height, against which ETDA-Net is robust, outperforming traditional deep learning based fall-detection methods.

Establishing diagnostic criteria for Perry syndrome.

OBJECTIVE: To establish international diagnostic criteria for Perry syndrome, a disorder characterised by clinical signs of parkinsonism, depression/apathy, weight loss, respiratory symptoms, mutations in the DCTN1 gene and TAR DNA-binding protein 43 (TDP-43) pathology. METHODS: Data from the published literature and newly identified patients were gathered and analysed during and after the International Symposium on Perry syndrome in Tokyo to identify diagnostic criteria for Perry syndrome. RESULTS: Eighty-seven patients with Perry syndrome carrying DCTN1 mutations from 20 families were included in this study, and common signs of the disorder were identified, including parkinsonism (95.2% of patients), depression/apathy (71.4%), respiratory symptoms (66.7%) and weight loss (49.2%). CONCLUSIONS: Based on our findings, we propose the following definitive diagnostic criteria for Perry syndrome: the presence of four cardinal signs of Perry syndrome, accompanied by a mutation in DCTN1; or a family history of the disease, parkinsonism and a mutation in DCTN1; or the presence of four cardinal signs and pathological findings that include nigral neuronal loss and TDP-43 pathology. As patients with Perry syndrome present with uniform clinical, genetic and pathological features, we further propose the disorder be termed 'Perry disease.'

Mutation screening of PLA2G6 in Japanese patients with early onset dystonia-parkinsonism.

A recessive mutation in PLA2G6, which is known to cause infantile neuroaxonal dystrophy (INAD) and neurodegeneration associated with brain iron accumulation (NBIA), has recently been shown to be responsible for PARK14-linked dystonia-parkinsonism. To study the frequency of PLA2G6 mutations, including those caused by gene rearrangement in patients with parkinsonism, we performed direct sequencing and investigated copy number variations (CNVs) of this gene in 109 Japanese patients with parkinsonism. Direct sequencing revealed a homozygous mutation (c.1495G>A; p.A499T), which is likely to be pathogenic and is already registered as rs141045127, and two compound-heterozygous mutations we have previously reported. No CNVs in PLA2G6 were detected in our subjects. Our results suggest that CNV in PLA2G6 is rare in parkinsonism, at least in the Japanese population, in contrast to the reports of its frequency in INAD. Further large studies in various populations are warranted to elucidate what causes the difference in frequencies of PLA2G6 rearrangement mutations between INAD and dystonia-parkinsonism.

A preliminary diffusional kurtosis imaging study of Parkinson disease: comparison with conventional diffusion tensor imaging.

INTRODUCTION: Diffusional kurtosis imaging (DKI) is a more sensitive technique than conventional diffusion tensor imaging (DTI) for assessing tissue microstructure. In particular, it quantifies the microstructural integrity of white matter, even in the presence of crossing fibers. The aim of this preliminary study was to compare how DKI and DTI show white matter alterations in Parkinson disease (PD). METHODS: DKI scans were obtained with a 3-T magnetic resonance imager from 12 patients with PD and 10 healthy controls matched by age and sex. Tract-based spatial statistics were used to compare the mean kurtosis (MK), mean diffusivity (MD), and fractional anisotropy (FA) maps of the PD patient group and the control group. In addition, a region-of-interest analysis was performed for the area of the posterior corona radiata and superior longitudinal fasciculus (SLF) fiber crossing. RESULTS: FA values in the frontal white matter were significantly lower in PD patients than in healthy controls. Reductions in MK occurred more extensively throughout the brain: in addition to frontal white matter, MK was lower in the parietal, occipital, and right temporal white matter. The MK value of the area of the posterior corona radiata and SLF fiber crossing was also lower in the PD group. CONCLUSION: DKI detects changes in the cerebral white matter of PD patients more sensitively than conventional DTI. In addition, DKI is useful for evaluating crossing fibers. By providing a sensitive index of brain pathology in PD, DKI may enable improved monitoring of disease progression.

Mechanisms of genomic instabilities underlying two common fragile-site-associated loci, PARK2 and DMD, in germ cell and cancer cell lines.

Common fragile sites (CFSs) are specific chromosome regions that exhibit an increased frequency of breaks when cells are exposed to a DNA-replication inhibitor such as aphidicolin. PARK2 and DMD, the causative genes for autosomal-recessive juvenile Parkinsonism and Duchenne and Becker muscular dystrophy, respectively, are two very large genes that are located within aphidicolin-induced CFSs. Gross rearrangements within these two genes are frequently observed as the causative mutations for these diseases, and similar alterations within the large fragile sites that surround these genes are frequently observed in cancer cells. To elucidate the molecular mechanisms underlying this fragility, we performed a custom-designed high-density comparative genomic hybridization analysis to determine the junction sequences of approximately 500 breakpoints in germ cell lines and cancer cell lines involving PARK2 or DMD. The sequence signatures where these breakpoints occur share some similar features both in germ cell lines and in cancer cell lines. Detailed analyses of these structures revealed that microhomologies are predominantly involved in rearrangement processes. Furthermore, breakpoint-clustering regions coincide with the latest-replicating region and with large nuclear-lamina-associated domains and are flanked by the highest-flexibility peaks and R/G band boundaries, suggesting that factors affecting replication timing collectively contribute to the vulnerability for rearrangement in both germ cell and somatic cell lines.

A case of α-synuclein gene duplication presenting with head-shaking movements.

BACKGROUND: PARK4 is a candidate locus for familial Parkinson's disease (PD), combined with multiplication of the α-synuclein gene (SNCA). The eventual phenotype is dependent on the copy number of SNCA. Mutations in leucine-rich repeat kinase 2 (LRRK2) are also causative of parkinsonism. This report describes a man who presented at our hospital complaining of a stagger after running and difficulty in handling the mouse of a personal computer, having suffered tremors since his twenties. Nine months after treatment and discharge, he developed titubation and began to drag his right foot. METHODS: We examined the patient's family pedigree for SNCA dosage, using quantitative polymerase chain reaction. We also screened this pedigree for mutations in parkin and LRRK2, using gene-sequencing techniques. RESULTS: We identified the proband, his sister, and his paternal uncle as carrying a duplication of SNCA. In addition, we found that the proband and his mother carried the G2385R variant of the LRRK2, a strong risk factor for PD in Asians and the rare V1450I variant, although only the proband showed symptoms of parkinsonism. No mutations were found in parkin. CONCLUSIONS: The combination of SNCA gene duplication and LRRK2 G2385R variant may explain the early onset of disease in this patient.

Population-specific frequencies for LRRK2 susceptibility variants in the Genetic Epidemiology of Parkinson's Disease (GEO-PD) Consortium.

BACKGROUND: Variants within the leucine-rich repeat kinase 2 gene are recognized as the most frequent genetic cause of Parkinson's disease. Leucine-rich repeat kinase 2 variation related to disease susceptibility displays many features that reflect the nature of complex, late-onset sporadic disorders like Parkinson's disease. METHODS: The Genetic Epidemiology of Parkinson's Disease Consortium recently performed the largest genetic association study for variants in the leucine-rich repeat kinase 2 gene across 23 different sites in 15 countries. RESULTS: Herein, we detail the allele frequencies for the novel risk factors (p.A419V and p.M1646T) and the protective haplotype (p.N551K-R1398H-K1423K) nominated in the original publication. Simple population allele frequencies not only can provide insight into the clinical relevance of specific variants but also can help genetically define patient groups. CONCLUSIONS: Establishing individual patient-based genomic susceptibility profiles that incorporate both risk factors and protective factors will determine future diagnostic and treatment strategies.

Relationship between cognitive impairment and white-matter alteration in Parkinson's disease with dementia: tract-based spatial statistics and tract-specific analysis.

OBJECTIVES: We investigated the relationship between white-matter alteration and cognitive status in Parkinson's disease (PD) with and without dementia by using diffusion tensor imaging. METHODS: Twenty PD patients, 20 PDD (Parkinson's disease with dementia) patients and 20 age-matched healthy controls underwent diffusion tensor imaging. The mean diffusivity and fractional anisotropy (FA) map of each patient group were compared with those of the control group by using tract-based spatial statistics. Tractography images of the genu of the corpus callosum fibre tracts were generated, and mean diffusivity and FA were measured. RESULTS: FA values in many major tracts were significantly lower in PDD patients than in control subjects; in the prefrontal white matter and the genu of the corpus callosum they were significantly lower in PDD patients than in PD patients. There was a significant correlation between the Mini-Mental State Examination (MMSE) scores and the FA values of the prefrontal white matter and the genu of the corpus callosum in patients with PD. CONCLUSIONS: Our study shows a relationship between cognitive impairment and alteration of the prefrontal white matter and genu of the corpus callosum. These changes may be useful in assessing the onset of dementia in PD patients. KEY POINTS: • Dementia is a common and important non-motor sign of Parkinson's disease (PD). • The neuropathological basis of dementia in PD is not clear. • DTI shows abnormalities in the prefrontal white matter in PD with dementia. • Prefrontal white matter alteration may be useful biomarker of dementia in PD.

Parkinson's disease-related LRRK2 G2019S mutation results from independent mutational events in humans.

Mutations in the leucine-rich-repeat kinase 2 (LRRK2) gene have been identified in families with autosomal dominant Parkinson's disease (PD) and in sporadic cases; the G2019S mutation is the single most frequent. Intriguingly, the frequency of this mutation in PD patients varies greatly among ethnic groups and geographic origins: it is present at <0.1% in East Asia, approximately 2% in European-descent patients and can reach frequencies of up to 15-40% in PD Ashkenazi Jews and North African Arabs. To ascertain the evolutionary dynamics of the G2019S mutation in different populations, we genotyped 74 markers spanning a 16 Mb genomic region around G2019S, in 191 individuals carrying the mutation from 126 families of different origins. Sixty-seven families were of North-African Arab origin, 18 were of North/Western European descent, 37 were of Jewish origin, mostly from Eastern Europe, one was from Japan, one from Turkey and two were of mixed origins. We found the G2019S mutation on three different haplotypes. Network analyses of the three carrier haplotypes showed that G2019S arose independently at least twice in humans. In addition, the population distribution of the intra-allelic diversity of the most widespread carrier haplotype, together with estimations of the age of G2019S determined by two different methods, suggests that one of the founding G2019S mutational events occurred in the Near East at least 4000 years ago.

Pseudo-heterozygous rearrangement mutation of parkin.

BACKGROUND: Mutations in parkin are the most frequent cause of autosomal recessive parkinsonism. Quantitative PCR is used to detect parkin rearrangements. However, the method has an inherent problem-deletion and duplication in the same allelic exon could be determined as normal. To present this misidentification, we report a family with compound heterozygous rearrangements in parkin. METHODS: A patient with early-onset parkinsonism and the parents were investigated by quantitative PCR, haplotype analysis, reverse-transcription PCR, and direct sequencing. RESULTS: A single heterozygous duplication (duplication of exons 6-7) was identified in the patient by quantitative PCR. Detailed analysis of the family revealed the patient carried compound heterozygous of combined deletion (deletion of exons 3-5) and duplication (duplication of exons 3-7). CONCLUSIONS: For correct determination of rearrangement mutation, mutation analysis of the patient as well as other family members and/or break-point analysis of genomic DNA and at the transcript level should be conducted.

VPS35 mutation in Japanese patients with typical Parkinson's disease.

Vacuolar protein sorting 35 (VPS35) was recently reported to be a pathogenic gene for late-onset autosomal dominant Parkinson's disease (PD), using exome sequencing. To date, VPS35 mutations have been detected only in whites with PD. The aim of the present study was to determine the incidence and clinical features of Asian PD patients with VPS35 mutations. We screened 7 reported nonsynonymous missense variants of VPS35, including p.D620N, known as potentially disease-associated variants of PD, in 300 Japanese index patients with autosomal dominant PD and 433 patients with sporadic PD (SPD) by direct sequencing or high-resolution melting (HRM) analysis. In addition, we screened 579 controls for the p.D620N mutation by HRM analysis. The p.D620N mutation was detected in 3 patients with autosomal dominant PD (1.0%), in 1 patient with SPD (0.23%), and in no controls. None of the other reported variants of VPS35 were detected. Haplotype analysis suggested at least 3 independent founders for Japanese patients with p.D620N mutation. Patients with the VPS35 mutation showed typical tremor-predominant PD. We report Asian PD patients with the VPS35 mutation. Although VPS35 mutations are uncommon in PD, the frequency of such mutation is relatively higher in Japanese than reported in other populations. In VPS35, p.D620N substitution may be a mutational hot spot across different ethnic populations. Based on the clinical features, VPS35 should be analyzed in patients with PD, especially autosomal dominant PD or tremor-predominant PD.

Lack of genetic association of the UCHL1 gene with Alzheimer's disease and Parkinson's disease with dementia.

BACKGROUND/AIMS: Several candidate genes were suggested to modify the susceptibility to both Alzheimer's disease (AD) and Parkinson's disease (PD). Symptoms of dementia are found in approximately 30% of PD patients. Both apolipoprotein E (APO E) and ubiquitin carboxyl-terminal esterase L1 (UCHL1) are neuropathogenic proteins for both diseases. The aim of this study was to investigate whether polymorphisms of both genes are associated with AD and PD with dementia (PDD). METHODS: The APO E polymorphism and 5 common single-nucleotide polymorphisms (SNPs) of the UCHL1 gene were analyzed using a case-control study design. RESULTS: Although APO E4 affected the onset of AD, the 5 SNPs of the UCHL1 gene were not associated with risk for AD. Linkage disequilibrium (LD) analysis of our Japanese data set showed that the SNPs of the UCHL1 gene are part of one LD block. Although one SNP, rs4861387, of the UCHL1 gene showed marginal association with PDD, we did not detect any association between the other SNPs and PDD. CONCLUSION: The common SNPs of UCHL1 are not major risk factors for AD. Since our analyses on PDD are preliminary, further genetic studies on APO E, UCHL1 and PD with and without dementia are needed.

Minimization of MEDA Biochip-Size in Droplet Routing.

With the increasing demand for fast, accurate, and reliable biological sensor systems, miniaturized systems have been aimed at droplet-based sensor systems and have been promising. A micro-electrode dot array (MEDA) biochip, which is one kind of the miniaturized systems for biochemical protocols such as dispensing, dilutions, mixing, and so on, has become widespread due to enabling dynamical control of the droplets in microfluidic manipulations. In MEDA biochips, the electrowetting-on-dielectric (EWOD) technique stands out since it can actuate droplets with nano/picoliter volumes. Microelectrode cells on MEDA actuate multiple droplets simultaneously to route locations for the purpose of the biochemical operations. Taking advantage of the feature, droplets are often routed in parallel to achieve high-throughput outcomes. Regarding parallel manipulation of multiple droplets, however, the droplets are known to be initially placed at a distant position to avoid undesirable mixing. The droplets thus result in traveling a long way for a manipulation, and the required biochip size for routing is also enlarged. This paper proposes a routing method for droplets to reduce the biochip size on a MEDA biochip with the allowance of splitting during routing operations. We mathematically derive the routing problem, and the experiments demonstrate that our proposal can significantly reduce the biochip size by 70.8% on average, compared to the state-of-the-art method.

Optimizing the Deep Neural Networks by Layer-Wise Refined Pruning and the Acceleration on FPGA.

To accelerate the practical applications of artificial intelligence, this paper proposes a high efficient layer-wise refined pruning method for deep neural networks at the software level and accelerates the inference process at the hardware level on a field-programmable gate array (FPGA). The refined pruning operation is based on the channel-wise importance indexes of each layer and the layer-wise input sparsity of convolutional layers. The method utilizes the characteristics of the native networks without introducing any extra workloads to the training phase. In addition, the operation is easy to be extended to various state-of-the-art deep neural networks. The effectiveness of the method is verified on ResNet architecture and VGG networks in terms of dataset CIFAR10, CIFAR100, and ImageNet100. Experimental results show that in terms of ResNet50 on CIFAR10 and ResNet101 on CIFAR100, more than 85% of parameters and Floating-Point Operations are pruned with only 0.35% and 0.40% accuracy loss, respectively. As for the VGG network, 87.05% of parameters and 75.78% of Floating-Point Operations are pruned with only 0.74% accuracy loss for VGG13BN on CIFAR10. Furthermore, we accelerate the networks at the hardware level on the FPGA platform by utilizing the tool Vitis AI. For two threads mode in FPGA, the throughput/fps of the pruned VGG13BN and ResNet101 achieves 151.99 fps and 124.31 fps, respectively, and the pruned networks achieve about 4.3× and 1.8× speed up for VGG13BN and ResNet101, respectively, compared with the original networks on FPGA.

PLA2G6 variant in Parkinson's disease.

PLA2G6 was reported recently as the causative gene for PARK14-linked autosomal recessive early-onset dystonia-parkinsonism. In a recent study in Singapore, heterozygous PLA2G6 p.P806R (c.2417C>G) mutation in exon 17 was reported to be a possible Parkinson's disease (PD)-related mutation. To determine the significance of the PLA2G6 mutation, we conducted an association study by performing direct sequencing of PLA2G6 exon 17 in 379 Japanese sporadic PD patients and 310 controls in the Japanese general population. In this group, we found 12 patients (12/379=3.16%) and 10 controls (10/310=3.23%) with a heterozygous p.P806R mutation (P=0.96, χ(2)=0.0019). Therefore, our large case-controlled study suggests that PLA2G6 p.P806R is not a disease-associated polymorphism in PD. Moreover, we performed direct sequencing of all exons and exon-intron boundaries of PLA2G6 in 116 Japanese patients with sporadic PD. Two single heterozygous variants (p.R301C or p.D331N) were found (both frequencies: 1/379 patients vs 0/310 controls) and the roles of their variants were unclear. Finally, combined with the previous report, our findings emphasize that PLA2G6 mutations are unlikely to be the major causes or risk factors of PD at least in Asian populations. However, further large studies in various populations are needed because patients with PLA2G6 mutations can show heterogeneous clinical features.

Comprehensive mutational analysis of LRRK2 reveals variants supporting association with autosomal dominant Parkinson's disease.

Parkinson's disease (PD) is a neurodegenerative disorder characterized by neurodegeneration, most notably of dopaminergic neurons in the substantia nigra. To date, six causative genes have been identified including LRRK2, whose mutations are the most frequent in autosomal dominant PD (Ad-PD). We conducted a comprehensive mutational analysis of LRRK2 in 30 Ad-PD (11 Japanese and 19 Caucasian) families employing a DNA microarray-based resequencing system and direct nucleotide sequence analysis, and identified 23 variants including two known mutations, p.G2019S and p.I1371V, in three Caucasian families and one Caucasian family, respectively, a novel putative pathogenic mutation, p.N1221K, in one Japanese family, and a known nonsynonymous variant, p.G2385R, in two Japanese families. Detailed analysis of the frequency of p.G2385R among 100 Japanese Ad-PD, 73 sporadic PD (sPD) and 238 controls revealed that the frequency of the p.G2385R variant was significantly higher in Ad-PD than in controls (allele frequency, 9.0 vs 2.1%) (χ(2)=16.32, P=5.34 × 10(-5)). The p.G2385R variant, however, did not show complete cosegregation with PD. In addition, the frequency of p.G2385R was also higher in sPD than in controls, although not significant (allele frequency, 3.4 vs 2.1%) (χ(2)=0.76, P=0.38). These observations support the possibility that p.G2385R is associated with an increased risk of PD.

Clinical course of the first Asian family with Parkinsonism related to SNCA triplication.

Triplication of SNCA is a rare cause of familial Parkinson's disease compared with duplication. Its clinical course is believed to be more robust than duplication, though it is uncertain. Marked as the first among the Asian population, we identified a Japanese family (paternal grandfather, father, and son) with SNCA triplication based on genetic and clinical analyses. The proband had a completely triplicated region including SNCA. This allele did not share any common haplotypes with those of previously reported Japanese families with SNCA duplication. Clinical analysis indicated early onset, rapidly progressive parkinsonism with mild levodopa response. Further studies are needed to clarify the gene dose effect of SNCA.

Rapid screening of ATP13A2 variant with high-resolution melting analysis.

Several genetic and environmental factors are involved in the pathogenesis of Parkinson's disease (PD). Recently, a novel variant of ATP13A2 (p.A746T) responsible for PARK9 was reported as a risk factor for PD in the Han-Chinese population. To investigate the role of this variant in Japanese PD patients, we examined 917 Japanese PD patients (871 index cases) and 190 controls by high-resolution melting curve analysis. We detected heterozygous p.A746T variant in a single patient with sporadic PD and a single control subject. These results suggest that ATP13A2 p.A746T variant is unlikely to play a role as a common risk factor or a pathogenic mutation for PD at least in Japanese. Our data on Japanese differ from those reported recently on Han-Chinese. Further studies are needed to confirm conclusions on roles of ATP13A2 variant in Asians or other populations.