An efficient ranking-based ensembled multiclassifier for neurodegenerative diseases classification using deep learning.

Neurodegenerative diseases are group of debilitating and progressive disorders that primarily affect the structure and functions of nervous system, leading to gradual loss of neurons and subsequent decline in cognitive, and behavioral activities. The two frequent diseases affecting the world's significant population falling in the above category are Alzheimer's disease (AD) and Parkinson's disease (PD). These disorders substantially impact the quality of life and burden healthcare systems and society. The demographic characteristics, and machine learning approaches have now been employed to diagnose these illnesses; however, they possess accuracy limitations. Therefore, the authors have developed ranking-based ensemble approach based on the weighted strategy of deep learning classifiers. The whole modeling procedure of the proposed approach incorporates three phases. In phase I, preprocessing techniques are applied to clean the noise in datasets to make it standardized according to deep learning models as it significantly impacts their performance. In phase II, five deep learning models are selected for classification and calculation of prediction results. In phase III, a ranking-based ensemble approach is proposed to ensemble the results of the five models after calculating the ranks and weights of them. In addition, the Magnetic Resonance Imaging (MRI) datasets named Alzheimer's Disease Neuroimaging Initiative (ADNI) for AD classification and Parkinson's Progressive Marker Initiative (PPMI) for PD classification are selected to validate the proposed approach. Furthermore, the proposed method achieved the classification accuracy on AD- Cognitive Normals (CN) at 97.89%, AD- Mild Cognitive Impairment (MCI) at 99.33% and CN-MCI at 99.44% and on PD-CN at 99.22%, PD- Scans Without Evidence of Dopaminergic Effect (SWEDD) at 97.56% and CN-SWEDD at 98.22% respectively. Also, the multi-class classification shows the promising accuracy of 97.18% for AD and 97.85% for PD for the proposed framework. The findings of the study show that the proposed deep learning-based ensemble technique is competitive for AD and PD prediction in both multiclass and binary class classification. Furthermore, the proposed approach enhances generalization performance in diagnosing neurodegenerative diseases and performs better than existing approaches.

Phosphorylation-dependent control of Activity-regulated cytoskeleton-associated protein (Arc) protein by TNIK.

Activity-regulated cytoskeleton-associated protein (Arc) is an immediate early gene product that support neuroplastic changes important for cognitive function and memory formation. As a protein with homology to the retroviral Gag protein, a particular characteristic of Arc is its capacity to self-assemble into virus-like capsids that can package mRNAs and transfer those transcripts to other cells. Although a lot has been uncovered about the contributions of Arc to neuron biology and behavior, very little is known about how different functions of Arc are coordinately regulated both temporally and spatially in neurons. The answer to this question we hypothesized must involve the occurrence of different protein post-translational modifications acting to confer specificity. In this study, we used mass spectrometry and sequence prediction strategies to map novel Arc phosphorylation sites. Our approach led us to recognize serine 67 (S67) and threonine 278 (T278) as residues that can be modified by TNIK, which is a kinase abundantly expressed in neurons that shares many functional overlaps with Arc and has, along with its interacting proteins such as the NMDA receptor, and been implicated as a risk factor for psychiatric disorders. Furthermore, characterization of each residue using site-directed mutagenesis to create S67 and T278 mutant variants revealed that TNIK action at those amino acids can strongly influence Arc's subcellular distribution and self-assembly as capsids. Together, our findings reveal an unsuspected connection between Arc and TNIK. Better understanding of the interplay between these two proteins in neuronal cells could lead to new insights about apparition and progression of psychiatric disorders. Cover Image for this issue: https://doi.org/10.1111/jnc.15077.

A novel fusion based convolutional neural network approach for classification of COVID-19 from chest X-ray images.

Coronavirus disease is a viral infection caused by a novel coronavirus (CoV) which was first identified in the city of Wuhan, China somewhere in the early December 2019. It affects the human respiratory system by causing respiratory infections with symptoms (mild to severe) like fever, cough, and weakness but can further lead to other serious diseases and has resulted in millions of deaths until now. Therefore, an accurate diagnosis for such types of diseases is highly needful for the current healthcare system. In this paper, a state of the art deep learning method is described. We propose COVDC-Net, a Deep Convolutional Network-based classification method which is capable of identifying SARS-CoV-2 infected amongst healthy and/or pneumonia patients from their chest X-ray images. The proposed method uses two modified pre-trained models (on ImageNet) namely MobileNetV2 and VGG16 without their classifier layers and fuses the two models using the Confidence fusion method to achieve better classification accuracy on the two currently publicly available datasets. It is observed through exhaustive experiments that the proposed method achieved an overall classification accuracy of 96.48% for 3-class (COVID-19, Normal and Pneumonia) classification tasks. For 4-class classification (COVID-19, Normal, Pneumonia Viral, and Pneumonia Bacterial) COVDC-Net method delivered 90.22% accuracy. The experimental results demonstrate that the proposed COVDC-Net method has shown better overall classification accuracy as compared to the existing deep learning methods proposed for the same task in the current COVID-19 pandemic.

The in vitro structure and functions of the disordered late embryogenesis abundant three proteins.

Late embryogenesis abundant (LEA) proteins are produced during seed embryogenesis and in vegetative tissue in response to various abiotic stressors. A correlation has been established between LEA expression and stress tolerance, yet their precise biochemical mechanism remains elusive. LEA proteins are very rich in hydrophilic amino acids, and they have been found to be intrinsically disordered proteins (IDPs) in vitro. Here, we perform biochemical and structural analyses of the four LEA3 proteins from Arabidopsis thaliana (AtLEA3). We show that the LEA3 proteins are disordered in solution but have regions with propensity for order. All LEA3 proteins were effective cryoprotectants of LDH in the freeze/thaw assays, while only one member, AtLEA3-4, was shown to bind Cu2+ and Fe3+ ions with micromolar affinity. As well, only AtLEA3-4 showed binding and a gain in α-helicity in the presence of the membrane mimic dodecylphosphocholine (DPC). We explored this interaction in greater detail using 15 N-heteronuclear single quantum coherence (HSQC) nuclear magnetic resonance, and demonstrate that two sets of conserved motifs present in AtLEA3-4 are involved in the interaction with the DPC micelles, which themselves gain α-helical structure.

Expression and Purification of an Intrinsically Disordered Protein.

Intrinsically disordered proteins (IDPs) describe a group of proteins that do not have a regular tertiary structure and typically have very little ordered secondary structure. Despite not following the biochemical dogma of "structure determines function" and "function determines structure," IDPs have been identified as having numerous biological functions. We describe here the steps to express and purify the intrinsically disordered stress response protein, Late embryogenesis abundant protein 3-2 from Arabidopsis thaliana (AtLEA 3-2), with 15N and 13C isotopes in E. coli, although the protocol can be adapted for any IDP with or without isotopic labeling. The atlea 3-2 gene has been cloned into the pET-SUMO vector that in addition to the SUMO portion encodes an N-terminal hexahistidine sequence (His-tag). This vector allows for the SUMO-AtLEA 3-2 fusion protein to be purified using Ni-affinity chromatography and, through the use of ubiquitin-like-specific protease 1 (Ulp1, a SUMO protease), results in an AtLEA 3-2 with a native N-terminus. We also describe the expression and purification of Ulp1 itself.

Conserved sequence motifs in the abiotic stress response protein late embryogenesis abundant 3.

LEA3 proteins, a family of abiotic stress proteins, are defined by the presence of a tryptophan-containing motif, which we name the W-motif. We use Pfam LEA3 sequences to search the Phytozome database to create a W-motif definition and a LEA3 sequence dataset. A comprehensive analysis of these sequences revealed four N-terminal motifs, as well as two previously undiscovered C-terminal motifs that contain conserved acidic and hydrophobic residues. The general architecture of the LEA3 sequences consisted of an N-terminal motif with a potential mitochondrial transport signal and the twin-arginine motif cut-site, followed by a W-motif and often a C-terminal motif. Analysis of species distribution of the motifs showed that one architecture was found exclusively in Commelinids, while two were distributed fairly evenly over all species. The physiochemical properties of the different architectures showed clustering in a relatively narrow range compared to the previously studied dehydrins. The evolutionary analysis revealed that the different sequences grouped into clades based on architecture, and that there appear to be at least two distinct groups of LEA3 proteins based on their architectures and physiochemical properties. The presence of LEA3 proteins in non-vascular plants but their absence in algae suggests that LEA3 may have arisen in the evolution of land plants.

A hybrid algorithm for speckle noise reduction of ultrasound images.

BACKGROUND AND OBJECTIVE: Medical images are contaminated by multiplicative speckle noise which significantly reduce the contrast of ultrasound images and creates a negative effect on various image interpretation tasks. In this paper, we proposed a hybrid denoising approach which collaborate the both local and nonlocal information in an efficient manner. The proposed hybrid algorithm consist of three stages in which at first stage the use of local statistics in the form of guided filter is used to reduce the effect of speckle noise initially. Then, an improved speckle reducing bilateral filter (SRBF) is developed to further reduce the speckle noise from the medical images. Finally, to reconstruct the diffused edges we have used the efficient post-processing technique which jointly considered the advantages of both bilateral and nonlocal mean (NLM) filter for the attenuation of speckle noise efficiently. METHODS: The performance of proposed hybrid algorithm is evaluated on synthetic, simulated and real ultrasound images. The experiments conducted on various test images demonstrate that our proposed hybrid approach outperforms the various traditional speckle reduction approaches included recently proposed NLM and optimized Bayesian-based NLM. RESULTS: The results of various quantitative, qualitative measures and by visual inspection of denoise synthetic and real ultrasound images demonstrate that the proposed hybrid algorithm have strong denoising capability and able to preserve the fine image details such as edge of a lesion better than previously developed methods for speckle noise reduction. CONCLUSIONS: The denoising and edge preserving capability of hybrid algorithm is far better than existing traditional and recently proposed speckle reduction (SR) filters. The success of proposed algorithm would help in building the lay foundation for inventing the hybrid algorithms for denoising of ultrasound images.

Genome Analysis of Conserved Dehydrin Motifs in Vascular Plants.

Dehydrins, a large family of abiotic stress proteins, are defined by the presence of a mostly conserved motif known as the K-segment, and may also contain two other conserved motifs known as the Y-segment and S-segment. Using the dehydrin literature, we developed a sequence motif definition of the K-segment, which we used to create a large dataset of dehydrin sequences by searching the Pfam00257 dehydrin dataset and the Phytozome 10 sequences of vascular plants. A comprehensive analysis of these sequences reveals that lysine residues are highly conserved in the K-segment, while the amino acid type is often conserved at other positions. Despite the Y-segment name, the central tyrosine is somewhat conserved, but can be substituted with two other small aromatic amino acids (phenylalanine or histidine). The S-segment contains a series of serine residues, but in some proteins is also preceded by a conserved LHR sequence. In many dehydrins containing all three of these motifs the S-segment is linked to the K-segment by a GXGGRRKK motif (where X can be any amino acid), suggesting a functional linkage between these two motifs. An analysis of the sequences shows that the dehydrin architecture and several biochemical properties (isoelectric point, molecular mass, and hydrophobicity score) are dependent on each other, and that some dehydrin architectures are overexpressed during certain abiotic stress, suggesting that they may be optimized for a specific abiotic stress while others are involved in all forms of dehydration stress (drought, cold, and salinity).