Biology subjects

Akiyama, M.

Publications and source records attributed to Akiyama, M..

At least 19 recordsLinked to original sources

bioRxiv2026

RIFT-VAE: grammar-conditioned pretraining and latent-space optimization for RNA inverse folding

BackgroundRNA inverse folding designs nucleotide sequences expected to adopt a prescribed secondary structure. Search-based solvers can optimize folding-model objectives effectively, but difficult targets can require extensive sampling, and structural optimization alone does not explicitly preserve the sequence distributions or conserved motifs of natural RNA families. MethodsWe developed RIFT-VAE, a Transformer-based conditional variational autoencoder that receives a context-free grammar parse-tree representation of a target secondary structure and generates nucleotide labels on the corresponding tree. The framework combines progressively richer grammar rules, self-refinement learning from generated structure-sequence pairs, and cross-entropy-method optimization in the learned latent space. We evaluated RNAfold minimum-free-energy agreement on an RNAcentral-derived test set and the EteRNA100 benchmark, compared the method with four search-based solvers under matched total time budgets, and examined GC-content control and covariance-model family annotation. ResultsThe complete pipeline achieved RNAfold-Correct/RNAfold-MCC values of 0.833/0.994 on the RNAcentral-derived test set and 0.760/0.977 on EteRNA100. Latent-space optimization accounted for the largest increase in exact structural recovery. Under a 3,600-s total budget on EteRNA100, sequences generated by RIFT-VAE improved the exact-match rate of every tested downstream search method when used as warm starts; the largest change was observed for RNAInverse (Correct, 0.297 to 0.803; MCC, 0.505 to 0.985). The pretrained model also produced sequences with measurable correct-family covariance-model hits and supported explicit GC-content conditioning. ConclusionsRIFT-VAE is best interpreted as a hybrid generative-search framework: pretraining supplies a structure- and family-informed proposal distribution, whereas latent optimization concentrates evaluations in high-scoring regions. The reported structural scores are specific to RNAfold minimum-free-energy validation and do not establish biochemical function. Orthogonal folding predictors, stricter homology-controlled splits, diversity-aware evaluation, architecture-matched dot-bracket ablations, and experimental assays remain priorities for validation.

bioinformatics↗

bioRxiv2026

Effects of Arachis hypogaea extract on TRPV4 activation and epidermal barrier function

Transient receptor potential vanilloid 4 (TRPV4) is a Ca2+-permeable non-selective cation channel and its activating stimuli include anandamide, bisandrographolide, citric acid, arachidonic acid metabolic products by epoxygenases, hypo-osmotic cell swelling, and warm temperature. TRPV4 is involved in Ca2+-dependent signal transduction in several tissues. Since the activation of TRPV4 facilitates adherens junction formation in the skin epithelium, compounds that activate TRPV4 are expected to maintain or improve the barrier function of epidermal cells. In this study, we found that the extract of Arachis hypogaea (A. hypogaea) activate human TRPV4 (hTRPV4). In the Ca2+-imaging experiment, the application of A. hypogaea extract exhibited an increase in intracellular Ca2+ concentration ([Ca2+]i) in HEK293T cells expressing hTRPV4. The [Ca2+]i increases by application of A. hypogaea extract were not observed in HEK293T cells expressing hTRPV1, mouse TRPV2, hTRPV3, hTRPM8, or hTRPA1. We then examined the physicochemical properties of the components responsible for TRPV4 activation. Ethanol extracts of A. hypogaea caused an increase in [Ca2+]i in hTRPV4-expressing HEK293JN cells, whereas water, chloroform, and hexane extracts showed no activity. Moreover, the application of A. hypogaea extract enhanced transepithelial electrical resistance in the keratinocyte monolayer. These results suggest that A. hypogaea extract may contribute to the maintenance and improvement of the epidermal barrier function.

pharmacology and toxicology↗

bioRxiv2025

Generative AI-driven artificial DNA design for enhancing inter-species gene activation and enzymatic degradation of PET

Conventional approaches to heterologous gene expression rely on codon optimization, which is limited to swapping synonymous codons and often fails to capture deeper adaptive changes. In contrast, naturally evolved orthologous genes between species often differ by more than just synonymous substitutions - they can include non-synonymous mutations, insertions, and deletions that confer functional adaptation to different host contexts. Here we present OrthologTransformer, a Transformer-based deep learning model that converts orthologous genes between species by learning from large-scale orthologous gene datasets curated for high-quality sequence alignments. The model recapitulates the full spectrum of evolutionary differences - from synonymous codon swaps to amino acid-changing mutations and indels - to predict a coding sequence optimized for a target species while preserving the proteins function. In extensive validation across diverse bacterial species pairs, OrthologTransformer significantly increased the conversion accuracy of generated genes to native target sequences compared to the original source genes, even for pairs with stark disparities in GC content and optimal growth temperature. The Transformers context-aware designs also favored conservative amino acid substitutions, maintaining protein functional integrity. As a proof-of-concept, an OrthologTransformer-designed PETase gene for Bacillus subtilis from the host species Ideonella sakaiensis was synthesized and expressed, yielding robust PET plastic-degradation activity that surpassed synonymous codon-optimized controls. These results establish OrthologTransformer as a powerful tool for de novo cross-species gene adaptation, transcending the limits of traditional codon optimization and enabling more effective heterologous gene performance in synthetic biology applications.

bioinformatics↗

bioRxiv2025

Class I myosins direct circumferential F-actin flows to define cell chirality

Eukaryotic cells possess intrinsic chirality in their structure, motility, and intracellular dynamics, which are designated cell chirality. Cell chirality participates in the left-right asymmetric morphogenesis and tissue integrity. However, the mechanisms of cell chirality formation remain elusive. In Drosophila, two evolutionarily conserved myosin I genes, Myosin 1D (Myo1D) and Myosin 1C (Myo1C), respectively, dictate the dextral and sinistral chirality of the cells and body. Here, we reported that Myo1D and Myo1C respectively directed the clockwise and counterclockwise circumferential flow of F-actin in Drosophila macrophages. Both induced the corresponding circular cytoplasm flows and depended on Myosin2 (Myo2). In a modified in vitro motility assay using near-physiological actin concentrations, Myo1D triggered the self-organization of the F-actin ring (chiral F-actin ring) that rotated clockwise; conversely, Myo1C induced the random flow of F-actin. The chiral F-actin ring implied that the F-actin bundle was parallelly and annularly polarized concerning its barbed pointed end. Considering that Myo1D and Myo1C are localized to the dorsal plasma membrane of macrophages, Myo1D and Myo1C might organize the parallelly polarized F-actin in macrophages. Our results suggest that Myo2 might drive the clockwise circumferential flow of F-actin along its parallel and annular polarity induced by Myo1D, which may be a molecular basis of cell and organ chirality.

cell biology↗

bioRxiv2024

Expression of elastin, FBXW2, fibrillin-1, and α-smooth muscle actin in damaged blood vessels of bovine legs

Elastic fibers present within the internal and external elastic laminae play a vital role in maintaining the health of blood vessels. The impairment of smooth muscle cell function results in the development of vascular diseases. In this study, blood vessels from bovine legs were carefully examined in culture dishes for 5 weeks. Examination of elastin, F-box and WD-40 domain-containing protein 2 (FBXW2), and fibrillin-1 in serial sections revealed that elastin and FBXW2 were present in the internal and external elastic laminae, whereas fibrillin-1 was present outside of them. After 5 weeks, elastin and FBXW2 maintained the shape of the elastic lamina, whereas the fibrillin-1-rich layer spread and migrated from the elastic lamina. When the blood vessel lumen was blocked by tissue rich in fibrillin-1, -smooth muscle actin (SMA) expression weakened. Although elastin and fibrillin-1 are components of elastic fibers, this study revealed that fibrillin-1 has a role different from that of elastin. This in vitro model is useful for identifying the mechanisms underlying vascular degradation.

cell biology↗

bioRxiv2023

Growth inhibitory factor/metallothionein-3 is a sulfane sulfur-binding protein

Cysteine-bound sulfane sulfur atoms in proteins have received much attention as key factors in cellular redox homeostasis. However, the role of sulfane sulfur in zinc regulation has been underinvestigated. We report here that cysteine-bound sulfane sulfur atoms serve as ligands to hold and release zinc ions in growth inhibitory factor (GIF)/metallothionein-3 (MT-3) with an unexpected C-S-S-Zn structure. Oxidation of such a zinc/persulfide cluster in Zn7GIF/MT-3 results in the release of zinc ions, and intramolecular tetrasulfide bridges in apo-GIF/MT-3 efficiently undergo S-S bond cleavage by thioredoxin to regenerate Zn7GIF/MT-3. Three-dimensional molecular modeling confirmed the critical role of the persulfide group in the thermostability and Zn-binding affinity of GIF/MT-3. The present discovery raises the fascinating possibility that the function of other Zn-binding proteins is controlled by sulfane sulfur.

biochemistry↗

bioRxiv2023

Characterizing the quality metric in genotype imputation

Large-scale imputation reference panels are now available and have contributed to efficient genome-wide association studies through genotype imputation. However, it is still under debate whether large-size multi-ancestry or small-size population-specific reference panels are the optimal choices for under-represented populations. We imputed genotypes of East Asian (EAS; 180k Japanese) subjects using the Trans-Omics for Precision Medicine (TOPMed) reference panel and found that the standard imputation quality metric (Rsq) substantially overestimated the dosage r2 (squared correlation between imputed dosage and true genotype). Variance component analysis of Rsq revealed that the increased imputed-genotype certainty (dosages closer to 0, 1, or 2) caused upward bias, indicating some systemic bias in the imputation. Through systematic simulations using different template switching rates ({theta} value) in the hidden Markov model, we uncovered that the lower {theta} value increased the imputed-genotype certainty and Rsq; however, dosage r2 was insensitive to the {theta} value, thereby causing a deviation. In simulated reference panels with different sizes and ancestral diversities, the {theta} value estimates from Minimac decreased with the size of a single ancestry and increased with the ancestral diversity. Thus, Rsq could overestimate or underestimate dosage r2 for a subpopulation in the multi-ancestry panel and the deviation represents different imputed-dosage distributions. Finally, despite the impact of {theta} value, distant ancestries in the reference panel contributed only a few additional variants passing a predefined Rsq threshold. We conclude that the {theta} value has a substantial impact on the imputed dosage and the imputation quality metric value.

bioinformatics↗

bioRxiv2022

Single-cell colocalization analysis using a deep generative model

1Analyzing colocalization of single cells with heterogeneous molecular phenotypes is essential for understanding cell-cell interactions, cellular responses to external stimuli, and their biological functions in diseases and tissues. However, high-throughput methods for identifying spatial proximity at single-cell resolution are practically unavailable. Here, we introduce DeepCOLOR, a computational framework based on a deep generative model that recovers inter-cellular colocalization networks with single cell resolution by the integration of single cell and spatial transcriptomes. It segregates cell populations defined by the colocalization relationships and predicts cell-cell interactions between colocalized single cells. DeepCOLOR could identify plausible cell-cell interaction candidates in mouse brain tissues, human squamous cell carcinoma samples, and human lung tissues infected with SARS-CoV-2 by reconstructing spatial colocalization maps at single-cell resolution. DeepCOLOR is typically applicable to studying cell-cell interactions in any spatial niche. Our newly developed computational framework could help uncover molecular pathways across single cells connected with colocalization networks.

bioinformatics↗

bioRxiv2022

Tiling mechanisms of the compound eye through geometrical tessellation

Tilling patterns are observed in many biological structures. Hexagonal tilling, commonly observed in the compound eyes of wild-type Drosophila, is dominant in nature; this dominance can probably be attributed to physical restrictions such as structural robustness, minimal boundary length, and space filling efficiency. Surprisingly, tetragonal tiling patterns are also observed in some Drosophila small eye mutants and aquatic crustaceans. Herein, geometrical tessellation is shown to determine the ommatidial tiling patterns. In small eye mutants, the hexagonal pattern is transformed into a tetragonal pattern as the relative positions of neighboring ommatidia are stretched along the dorsal-ventral axis. Hence, the regular distribution of ommatidia and their uniform growth collectively play an essential role in the establishment of tetragonal and hexagonal tiling patterns in compound eyes.

developmental biology↗

bioRxiv2021

A multi-layer functional genomic analysis to understand noncoding genetic variation in lipids

A major challenge of genome-wide association studies (GWAS) is to translate phenotypic associations into biological insights. Here, we integrate a large GWAS on blood lipids involving 1.6 million individuals from five ancestries with a wide array of functional genomic datasets to discover regulatory mechanisms underlying lipid associations. We first prioritize lipid-associated genes with expression quantitative trait locus (eQTL) colocalizations, and then add chromatin interaction data to narrow the search for functional genes. Polygenic enrichment analysis across 697 annotations from a host of tissues and cell types confirms the central role of the liver in lipid levels, and highlights the selective enrichment of adipose-specific chromatin marks in high-density lipoprotein cholesterol and triglycerides. Overlapping transcription factor (TF) binding sites with lipid-associated loci identifies TFs relevant in lipid biology. In addition, we present an integrative framework to prioritize causal variants at GWAS loci, producing a comprehensive list of candidate causal genes and variants with multiple layers of functional evidence. Two prioritized genes, CREBRF and RRBP1, show convergent evidence across functional datasets supporting their roles in lipid biology.

genomics↗

bioRxiv2021

Informative RNA-base embedding for functional RNA structural alignment and clustering by deep representation learning

Effective embedding is being actively conducted by applying deep learning to biomolecular information. Obtaining better embedding enhances the quality of downstream analysis such as DNA sequence motif detection and protein function prediction. In this study, we adopt a pre-training algorithm for the effective embedding of RNA bases to acquire semantically rich representations, and apply it to two fundamental RNA sequence problems: structural alignment and clustering. By using the pre-learning algorithm to embed the four bases of RNA in a position-dependent manner using a large number of RNA sequences from various RNA families, a context-sensitive embedding representation is obtained. As a result, not only base information but also secondary structure and context information of RNA sequences are embedded for each base. We call this "informative base embedding" and use it to achieve accuracy superior to that of existing state-of-the-art methods in RNA structural alignment and RNA family clustering tasks. Furthermore, by performing RNA sequence alignment combining this informative base embedding with a simple Needleman-Wunsch alignment algorithm, we succeed in calculating a structural alignment in a time complexity O(n2) instead of the O(n6) time complexity of Sankoff-style algorithms.

bioinformatics↗

bioRxiv2021

Genomic style: yet another deep-learning approach to characterize bacterial genome sequences

MotivationBiological sequence classification is the most fundamental task in bioinformatics analysis. For example, in metagenome analysis, binning is a typical type of DNA sequence classification. In order to classify sequences, it is necessary to define sequence features. The k-mer frequency, base composition, and alignment-based metrics are commonly used. In contrast, in the field of image recognition using machine learning, image classification is broadly divided into those based on shape and those based on style. A style matrix was introduced as a method of expressing the style of an image (e.g., color usage and texture). ResultsWe propose a novel sequence feature, called genomic style, inspired by image classification approaches, for classifying and clustering DNA sequences. As with the style of images, the DNA sequence is considered to have a genomic style unique to the bacterial species, and the style matrix concept is applied to the DNA sequence. Our main aim is to introduce the genomics style as yet another basic sequence feature for metagenome binning problem in replace of the most commonly used sequence feature k-mer frequency. Performance evaluations show that our method using style matrix achieves the superior accuracy than state-of-the-art binning tools based on k-mer frequency.

bioinformatics↗

bioRxiv2021

Multi-Ancestry Meta-Analysis yields novel genetic discoveries and ancestry-specific associations

We present a new method, Multi-Ancestry Meta-Analysis (MAMA), which combines genome-wide association study (GWAS) summary statistics from multiple populations to produce new summary statistics for each population, identifying novel loci that would not have been discovered in either set of GWAS summary statistics alone. In simulations, MAMA increases power with less bias and generally lower type-1 error rate than other multi-ancestry meta-analysis approaches. We apply MAMA to 23 phenotypes in East-Asian- and European-ancestry populations and find substantial gains in power. In an independent sample, novel genetic discoveries from MAMA replicate strongly.

genomics↗

bioRxiv2021

Genetically independent phenotype analysis identifies LPA and VCAM1 as drug targets for human ageing

The length and quality of life is important to us all, yet identification of promising drug targets for human ageing using genetics has had limited success. Here, we combine six large European-ancestry genome-wide association studies (GWAS) of human ageing traits--healthspan, father and mother lifespan, exceptional longevity, frailty index, and self-rated health--in a principal component framework that maximises their shared genetic architecture. The first principal component (GIP1) is more heritable than the original studies and shows strong genetic correlations with length of life as well as multiple indices of mental and physical wellbeing. We identify 27 genomic regions associated with GIP1, and provide additional, independent evidence for an effect on human ageing for loci near HTT and MAML3 using a study of Finnish and Japanese subject survival. Across the genome, GIP1 associations are enriched in genes involved in haem metabolism and pathways related to transcription, neurogenesis, homeostasis, proteolysis, intracellular signalling, immunity, and the muscle system. Finally, using proteome-wide two-sample Mendelian randomisation and colocalisation, we provide robust evidence for a detrimental effect of blood levels of apolipoprotein(a) (LPA) and vascular cell adhesion molecule 1 (VCAM1) on GIP1. Together, our results demonstrate that combining multiple ageing traits using genetic principal components enhances power to detect biological targets for human ageing.

genomics↗

bioRxiv2020

Role of FBXW2 in explant culture of bovine periosteum-derived cells

Osteoporosis and bone fracture decrease quality of life. Bone regeneration is a notable technique for osteoporosis treatment. A previous study reported that F-box and WD-40 domain-containing protein 2 (FBXW2) and osteocalcin have the same shape in the periosteum after 5 weeks. However, the osteoblastic functions of FBXW2 are not clear. In this study, double fluorescent immunostaining revealed a small amount of osteocalcin in the area of FBXW2 aggregation at 1 week, periosteal cells, and osteocalcin pushed toward the edge of periosteum, and, apart from FBXW2 tubes at 2 weeks, multilayered periosteum-derived cells at 3 weeks and sticking of osteocalcin in the periosteum with cells at 4 weeks. At 5 weeks, FBXW2 disappeared at the root of periosteum-derived cells, while osteocalcin and cells remained. Based on these results, it is hypothesized that FBXW2 maintains tissue shapes and prevents escape of inner periosteal cells, and the disappearance of FBXW2 causes migration of periosteum-derived cells out of the periosteum along with osteocalcin. Furthermore, FBXW2 may play a role in dynamic tissue remodeling and bone formation.

cell biology↗

bioRxiv2020

Curing genetic skin disease through altered replication stress response

Revertant mosaicism, or natural gene therapy, refers to the spontaneous in vivo reversion of an inherited mutation in a somatic cell1. Only [~]50 human genetic disorders exhibit revertant mosaicism, implicating a distinctive role played by mutant proteins in somatic correction of a pathogenic germline mutation2. However, the process by which mutant proteins induce somatic genetic reversion in these diseases remains unknown. Here we show that heterozygous pathogenic CARD14 mutations causing autoinflammatory skin diseases, including psoriasis and pityriasis rubra pilaris, are repaired mainly via homologous recombination. Rather than altering the DNA damage response to exogenous stimuli such as X-irradiation or etoposide treatment, mutant CARD14 increased DNA double-strand breaks under conditions of replication stress. Furthermore, mutant CARD14 suppressed new origin firings without promoting crossover events in the replication stress state. Together, these results suggest that mutant CARD14 alters the replication stress response and preferentially drives break-induced replication (BIR), which is generally suppressed in eukaryotes3. Our results highlight the involvement of BIR in reversion events, thus revealing a previously undescribed role of BIR that could potentially be exploited to develop therapeutics for currently intractable genetic diseases.

molecular biology↗

bioRxiv2020

RNA secondary structure prediction using deep learning with thermodynamic integrations

RNA secondary structure prediction is one of the key technologies for revealing the essential roles of functional non-coding RNAs. Although machine learning-based rich-parametrized models have achieved extremely high performance in terms of prediction accuracy, the risk of overfitting for such models has been reported. In this work, we propose a new algorithm for predicting RNA secondary structures that uses deep learning with thermodynamic integration, thereby enabling robust predictions. Similar to our previous work, the folding scores, which are computed by a deep neural network, are integrated with traditional thermodynamic parameters to enable robust predictions. We also propose thermodynamic regularization for training our model without overfitting it to the training data. Our algorithm (MXfold2) achieved the most robust and accurate predictions in computational experiments designed for newly discovered non-coding RNAs, with significant 2-10 % improvements over our previous algorithm (MXfold) and standard algorithms for predicting RNA secondary structures in terms of F-value.

bioinformatics↗

bioRxiv2019

Systematic detection of Mendelian and non-Mendelian variants associated with retinitis pigmentosa by genome-wide association study

To uncover genetic basis of autosomal recessive retinitis pigmentosa (ARRP), we applied 2-step genome-wide association study (GWAS) in 640 Japanese patients prescreened with targeted re-sequencing. Meta-GWAS identified three independent peaks at P < 5.0x10-8, all within the major ARRP gene EYS. Two were each tagged by a low frequency variant (allele frequency < 0.05); a known founder Mendelian mutation (c.4957dupA, p.S1653Kfs*2) and a presumably hypomorphic non-synonymous variant (c.2528G>A, p.G843E). c.2528G>A newly solved 7.0% of Japanese ARRP cases, improving genetic diagnosis by 26.8% and simultaneously serving as a new attractive target for genome editing gene therapy. The third peak was tagged by an intronic common variant, representing a novel disease-susceptibility signal. GWAS successfully unraveled genetic causes of a rare "monogenic" disorder for the first time, which provided unexpected insights into significant contribution of non-Mendelian genetic factors and identified a novel high frequency variant directly linked to development of local genome therapeutics.

genetics↗