Biology subjects

Tomii, K.

Publications and source records attributed to Tomii, K..

10 recordsLinked to original sources

bioRxiv2026

Prot-LAMBDA: Explicit Distance Learning Enhances Structural Reasoning in Protein Language Models

Protein language models (PLMs) learn evolutionary information from large-scale sequence data, but three-dimensional relationships are encoded only implicitly. Here, we introduce Prot-LAMBDA (Protein LAnguage Model Boosted with Distance Awareness), a PLM that explicitly incorporates spatial relationships by coupling residue embeddings with inter-residue contacts. Prot-LAMBDA improves performance across diverse structure-related tasks, including contact, secondary structure, backbone geometry, solvent accessibility, and protein fold prediction. Notably, it achieves a twofold improvement in long-range contact recall and an 11.7% reduction in {psi}-angle prediction error relative to ESM2-3B. Despite having approximately fivefold fewer parameters, Prot-LAMBDA also improves 3D structure prediction over ESM2-3B by 5-7% in TM-score when coupled to the same structure-prediction module. Building on these representations, we developed LambdaFold, a lightweight distance-guided structure prediction framework that achieves performance comparable to ESMFold on proteins strictly non-redundant to the training data. Finally, retrieval-augmented integration of structural templates increases mean TM-score substantially for targets with high template coverage and rescues several incorrect folds. Together, these results demonstrate that explicit spatial constraints enable efficient and generalizable structural representation learning and protein structure prediction.

bioinformatics↗

bioRxiv2026

zsasa: a Zig-based engine for high-throughput solvent accessible surface area at proteome scale

Solvent accessible surface area (SASA) is widely used to describe protein stability, ligand binding, mutation effects, and protein-protein interfaces. As structural biology workloads expand to predicted-structure col-lections, trajectories, and large assemblies, SASA tools must combine reproducible calculation with high throughput, low memory use, and workflow-friendly input handling. We present zsasa, a Zig-based SASA engine with command-line and Python interfaces. zsasa implements the established Shrake-Rupley and Lee-Richards algorithms, provides exact f64/f32 modes and an optional bitmask approximation, and supports batch and trajectory workflows, compressed structure inputs, and configurable atom classification including Chemical Component Dictionary (CCD)-based radii for non-standard components. In matched Shrake-Rupley validation on 4,370 Escherichia coli AlphaFold Database structures, exact double-precision zsasa reproduced FreeSASA total SASA values to near numerical identity. In 10-thread batch benchmarks on the E. coli and 23,586-structure human AlphaFold collections, zsasa achieved a 2.94-fold speedup over a FreeSASA batch wrapper in exact f64 mode. In bitmask mode, zsasa reached up to a 9.70-fold speedup, using roughly 12.5% to 25% of the comparator peak memory. Trajectory benchmarks exceeded 1,000 frames/s at tens of megabytes of peak memory, and a 4.5-million-atom PDB stress-test file completed in less than 5 s. These results support zsasa as a practical tool for reproducible, low-memory generation of surface-derived structural features at large scale. zsasa is available under the MIT License at https://github.com/N283T/zsasa.

bioinformatics↗

bioRxiv2025

LRRK2 is activated by phosphatidylinositol 3-phosphate in conjunction with CASM

LRRK2 is a Parkinsons disease (PD)-associated kinase that phosphorylates Rab GTPases. Hyperactivation of LRRK2, which is thought to cause PD, occurs on stressed endolysosomal membranes via CASM (conjugation of ATG8 to single membranes) and Rab, whereas the involvement of specific membrane lipids remained unclear. Here, we found that LRRK2 was potently activated upon treatment with PIKfyve inhibitors that increase cellular phosphatidylinositol 3-phosphate (PI3P) levels, and that additional treatment with the class III phosphoinositide 3-kinase (PI3K) inhibitors blocking PI3P generation suppressed LRRK2 activation. Cellular PI3P levels were indeed correlated with the LRRK2 kinase activity. CASM-induced LRRK2 activation, but not CASM itself, was suppressed by PI3K inhibitors, whereas inhibition of CASM or Rab12/Rab29 suppressed LRRK2 activation induced by PIKfyve inhibitors. LRRK2 and phospho-Rab10 were co-localized with PI3P on LAMP1-positive enlarged vacuoles, and LRRK2 directly bound PI3P in vitro via its N-terminal positively charged amino acids. Together, we propose an updated mechanism of LRRK2 activation that requires CASM, Rab and PI3P-rich membranes.

cell biology↗

bioRxiv2024

A transphyletic study of metazoan β-catenin protein complexes

{beta}-catenin is essential for various biological processes, such as body axis determination and cell differentiation, during embryonic development in metazoans. {beta}-catenin functions are thought to be exerted through complexes formed with various types of proteins. Although {beta}-catenin complex proteins have been identified in several bilaterians, little is known about the structural and functional properties of {beta}-catenin complexes in early metazoan evolution. In this study, we performed a comparative analysis of {beta}-catenin sequences in nonbilaterian lineages that diverged early in metazoan evolution. We also carried out transphyletic function experiments with {beta}-catenin from nonbilaterian metazoans using developing Xenopus embryos, which included secondary axis induction in embryos and proteomic analysis of {beta}-catenin protein complexes. Comparative functional analysis of nonbilaterian {beta}-catenins demonstrated sequence characteristics important for {beta}-catenin functions, and the deep origin and evolutionary conservation of the cadherin-catenin complex. Proteins co-immunoprecipitated with {beta}-catenin included several proteins conserved among metazoans. These data provide new insights into the conserved repertoire of {beta}-catenin complexes.

evolutionary biology↗

bioRxiv2024

Neuronal and non-neuronal functions of the synaptic cell adhesion molecule, neurexin, in Nematostella vectensis.

The transition from diffusion-mediated cell-cell communication to faster and more targeted synaptic signaling in animal nervous systems has long been of interest to evolutionary biologists. Although genome sequencing of early-diverging metazoans has revealed the broad distribution of synapse-related genes among them, synaptic structures have been observed only in Cnidaria, the sister group to Bilateria. How synaptic machinery evolved remains largely unknown. In this study, we investigated the function of neurexins (Nrxns), a core family of presynaptic cell adhesion molecules with critical roles in bilaterian chemical synapses, using the cnidarian model, Nematostella vectensis. Neural Nrxns, named delta-Nrxns, are expressed mainly in neuronal cell clusters that exhibit both peptidergic and classical neurotransmitter signaling. Knockdown of Nrxn{delta} genes reduced spontaneous peristalsis of N. vectensis polyps. Interestingly, gene knockdown and pharmacological studies suggested that Nrxn{delta} is involved in glutamate- and glycine-mediated signaling rather than peptidergic signaling. Knockdown of the epithelial Nrxn in N. vectensis revealed a major role in cell adhesion, particularly between ectodermal and endodermal epithelia. Overall, this study provides molecular, functional, and cellular insights into the ancestral, non-neural function of Nrxns, as well as key information for understanding how and why this family of cell adhesion molecules was recruited to synaptic machinery.

evolutionary biology↗

bioRxiv2022

Global observation of plankton communities from space

Satellite remote sensing from space is a powerful way to monitor the global dynamics of marine plankton. Previous research has focused on developing models to predict the size or taxonomic groups of phytoplankton. Here we present an approach to identify representative communities from a global plankton network that included both zooplankton and phytoplankton and using global satellite observations to predict their biogeography. Six representative plankton communities were identified from a global co-occurrence network inferred using a novel rDNA 18S V4 planetary-scale eukaryotic metabarcoding dataset. Machine learning techniques were then applied to train a model that predicted these representative communities from satellite data. The model showed an overall 67% accuracy in the prediction of the representative communities. The prediction based on 17 satellite-derived parameters showed better performance than based only on temperature and/or the concentration of chlorophyll a. The trained model allowed to predict the global spatiotemporal distribution of communities over 19-years. Our model exhibited strong seasonal changes in the community compositions in the subarctic-subtropical boundary regions, which were consistent with previous field observations. This network-oriented approach can easily be extended to more comprehensive models including prokaryotes as well as viruses.

ecology↗

bioRxiv2022

SARS-CoV-2 ORF8 is a viral cytokine regulating immune responses

Many patients with severe COVID-19 suffer from pneumonia, and thus elucidation of the mechanisms underlying the development of such severe pneumonia is important. The ORF8 protein is a secreted protein of SARS-CoV-2, whose in vivo function is not well understood. Here, we analyzed the function of ORF8 protein by generating ORF8-knockout SARS-CoV-2. We found that the lung inflammation observed in wild-type SARS-CoV-2-infected hamsters was decreased in ORF8-knockout SARS-CoV-2-infected hamsters. Administration of recombinant ORF8 protein to hamsters also induced lymphocyte infiltration into the lungs. Similar pro-inflammatory cytokine production was observed in primary human monocytes treated with recombinant ORF8 protein. Furthermore, we demonstrate that the serum ORF8 protein levels are correlated well with clinical markers of inflammation. These results demonstrated that the ORF8 protein is a viral cytokine of SARS-CoV-2 involved in the in the immune dysregulation observed in COVID-19 patients, and that the ORF8 protein could be a novel therapeutic target in severe COVID-19 patients.

microbiology↗

bioRxiv2020

An infectivity-enhancing site on the SARS-CoV-2 spike protein is targeted by COVID-19 patient antibodies

SARS-CoV-2 infection causes severe symptoms in a subset of patients, suggesting the presence of certain unknown risk factors. Although antibodies against the receptor-binding domain (RBD) of the SARS-CoV-2 spike have been shown prevent SARS-CoV-2 infection, the effects of antibodies against other spike protein domains are largely unknown. Here, we screened a series of anti-spike monoclonal antibodies from COVID-19 patients, and found that some of antibodies against the N-terminal domain (NTD) dramatically enhanced the binding capacity of the spike protein to ACE2, and thus increased SARS-CoV2 infectivity. Surprisingly, mutational analysis revealed that all the infectivity-enhancing antibodies recognized a specific site on the surface of the NTD. The antibodies against this infectivity-enhancing site were detected in all samples of hospitalized COVID-19 patients in the study. However, the ratio of infectivity-enhancing antibodies to neutralizing antibodies differed among patients. Furthermore, the antibodies against the infectivity-enhancing site were detected in 3 out of 48 uninfected donors, albeit at low levels. These findings suggest that the production of antibodies against SARS-CoV-2 infectivity-enhancing site could be considered as a possible exacerbating factors for COVID-19 and that a spike protein lacking such antibody epitopes may be required for safe vaccine development, especially for individuals with pre-existing enhancing antibodies.

microbiology↗

bioRxiv2018

Deep Neural Network for Protein Contact Prediction by Weighting Sequences in a Multiple Sequence Alignment

Protein contact prediction is a crucially important step for protein structure prediction. To predict a contact, approaches of two types are used: evolutionary coupling analysis (ECA) and supervised learning. ECA uses a large multiple sequence alignment (MSA) of homologue sequences and extract correlation information between residues. Supervised learning uses ECA analysis results as input features and can produce higher accuracy. As described herein, we present a new approach to contact prediction which can both extract correlation information and predict contacts in a supervised manner directly from MSA using a deep neural network (DNN). Using DNN, we can obtain higher accuracy than with earlier ECA methods. Simultaneously, we can weight each sequence in MSA to eliminate noise sequences automatically in a supervised way. It is expected that the combination of our method and other meta-learning methods can provide much higher accuracy of contact prediction.

bioinformatics↗

bioRxiv2016

Simple adjustment of the sequence weight algorithm remarkably enhances PSI-BLAST performance

PSI-BLAST, an extremely popular tool for sequence similarity search, features the utilization of Position Specific Scoring Matrix (PSSM) constructed from a multiple sequence alignment (MSA). PSSM allows the detection of more distant homologs than a general amino acid substitution matrix does. An accurate estimation of the weights of sequences in an MSA is crucially important for PSSM construction. PSI-BLAST divides a given MSA into multiple blocks, for which sequence weights are calculated. When the block width becomes very narrow, the sequence weight calculation can be difficult.\n\nWe demonstrate that PSI-BLAST indeed generates a significant fraction of blocks having widths less than 5, thereby degrading the PSI-BLAST performance. We revised the code of PSI-BLAST to prevent the blocks from being narrower than a given minimum block width (MBW). We designate the modified application of PSI-BLAST as PSI-BLASTexB. When MBW is 25, PSI-BLASTexB notably outperforms PSI-BLAST consistently for three independent benchmark sets. The performance boost is even more drastic when an MSA, instead of a sequence, was used as a query.\n\nOur results demonstrate that the generation of narrow-width blocks during the sequence weight calculation is a critically important factor that restricts the PSI-BLAST search performance. By preventing narrow blocks, PSI-BLASTexB remarkably upgrades the PSI-BLAST performance.

bioinformatics↗