Biology subjects

Mamitsuka, H.

Publications and source records attributed to Mamitsuka, H..

10 recordsLinked to original sources

bioRxiv2026

Improving Biological Sequence Prediction with AlphaFold2 Representation

MotivationAccurate prediction of functional sites from primary sequences is essential for elucidating biological mechanisms and advancing rational drug design. However, traditional sequence-based features are inherently unable to capture complex structural protein contexts. Recently, AlphaFold2 (AF2) revolutionized protein structure prediction, raising expectations of AF2 to serve as a feature extractor providing structure-rich representation, which can be useful for sequence-based prediction, particularly for unknown sequences. ResultsWe present a novel feature-engineering paradigm that leverages a high-dimensional latent representation matrix (of L x D, where L is the sequence length and D is the feature dimension size) extracted directly from the AF2 Evoformer module. We systematically evaluated the AF2 representation, comparing with conventional sequence-based features, such as hidden Markov model profiles, using a variety of machine learning models, on two structurally contrasting tasks, calpain cleavage site and nucleic-acid-binding site prediction. The AF2 representation outperformed conventional sequence-based features clearly and entirely, particularly for targets with low sequence homology to training data. Furthermore, interpretability analyses, using SHapley Additive exPlanations (SHAP) and Uniform MAnifoldapproximation and Projection (UMAP), showed more details behind the performance advantage of AF2 representation through feature importance ranking and visualization. Overall, these empirical results confirmed that AF2 representation could effectively bridge the sequence-to-structure gap as a feature input for sequence prediction, without increasing heavy computational burden. Availability and implementationSource code, pre-trained models, and datasets are freely available to non-commercial users at https://github.com/Lili-irtyd/Improve-biological-sequences-prediction-by-AlphaFold2. Contactmami@kuicr.kyoto-u.ac.jp

bioinformatics↗

bioRxiv2026

DeepKOALA: A Fast and Accurate Deep Learning Framework for KEGG Orthology Assignment

The KEGG Orthology (KO) system links DNA and protein sequences to biological functions and pathways, providing a curated, fundamental and consistent annotation framework across all domains of life. While accurate, traditional sequence alignment-based annotation methods are computationally expensive, which severely limits their application in large-scale datasets. To address this challenge, we introduce DeepKOALA, a deep learning approach based on Gated Recurrent Units (GRU), which frames KO annotation as an open-set recognition task. This design reduces false positives arising from out-of-scope sequences and, together with a lightweight GRU backbone, enables high-throughput annotation. The GRU-based model was benchmarked against four other deep learning architectures and showed the best balance between speed and accuracy. We further performed a cross-species evaluation of DeepKOALA in comparison with existing KO annotation tools, where it achieved an F1-score of 0.8653 and 37.5-fold acceleration compared with BlastKOALA. We also provide a specialized fragment model for handling incomplete sequences and an optional Multi-domain mode. Together, these features make DeepKOALA a scalable, efficient, and accurate solution for high-throughput function annotation. O_FIG O_LINKSMALLFIG WIDTH=200 HEIGHT=61 SRC="FIGDIR/small/698072v1_ufig1.gif" ALT="Figure 1"> View larger version (15K): org.highwire.dtl.DTLVardef@8c3b53org.highwire.dtl.DTLVardef@8acfa8org.highwire.dtl.DTLVardef@147398borg.highwire.dtl.DTLVardef@112d762_HPS_FORMAT_FIGEXP M_FIG C_FIG

bioinformatics↗

bioRxiv2021

DeepMHCII: A Novel Binding Core-Aware Deep InteractionModel for Accurate MHC II-peptide Binding Affinity Prediction

Computationally predicting MHC-peptide binding affinity is an important problem in immunological bioinformatics. Recent cutting-edge deep learning-based methods for this problem are unable to achieve satisfactory performance for MHC class II molecules. This is because such methods generate the input by simply concatenating the two given sequences: (the estimated binding core of) a peptide and (the pseudo sequence of) an MHC class II molecule, ignoring the biological knowledge behind the interactions of the two molecules. We thus propose a binding core-aware deep learning-based model, DeepMHCII, with binding interaction convolution layer (BICL), which allows integrating all potential binding cores (in a given peptide) and the MHC pseudo (binding) sequence, through modeling the interaction with multiple convolutional kernels. Extensive empirical experiments with four large-scale datasets demonstrate that DeepMHCII significantly outperformed four state-of-the-art methods under numerous settings, such as five-fold cross-validation, leave one molecule out, validation with independent testing sets, and binding core prediction. All these results with visualization of the predicted binding cores indicate the effectiveness and importance of properly modeling biological facts in deep learning for high performance and knowledge discovery. DeepMHCII is publicly available at https://weilab.sjtu.edu.cn/DeepMHCII/.

bioinformatics↗

bioRxiv2021

Crop loss identification at field parcel scale using satellite remote sensing and machine learning

Identifying crop loss at field parcel scale using satellite images is challenging: first, crop loss is caused by many factors during the growing season; second, reliable reference data about crop loss are lacking; third, there are many ways to define crop loss. This study investigates the feasibility of using satellite images to train machine learning (ML) models to classify agricultural field parcels into those with and without crop loss. The reference data for this study was provided by Finnish Food Authority (FFA) containing crop loss information of approximately 1.4 million field parcels in Finland covering about 3.5 million ha from 2000 to 2015. This reference data was combined with Normalised Difference Vegetation Index (NDVI) derived from Landsat 7 images, in which more than 80% of the possible data are missing. Despite the hard problem with extremely noisy data, among the four ML models we tested, random forest (with mean imputation and missing value indicators) achieved the average AUC (area under the ROC curve) of 0.688 {+/-} 0.059 over all 16 years with the range [0.602, 0.795] in identifying new crop-loss fields based on reference fields of the same year. To our knowledge, this is one of the first large scale benchmark study of using machine learning for crop loss classification at field parcel scale. The classification setting and trained models have numerous potential applications, for example, allowing government agencies or insurance companies to verify crop-loss claims by farmers and realise efficient agricultural monitoring.

bioinformatics↗

bioRxiv2020

XGSEA: CROSS-species Gene Set Enrichment Analysis via domain adaptation

Gene set enrichment analysis (GSEA) has been widely used to identify gene sets with statistically significant difference between cases and controls against a large gene set. GSEA needs both phenotype labels and expression of genes. However, gene expression are assessed more often for model organisms than minor species. More importantly, gene expression could not be measured under specific conditions for human, due to high healthy risk of direct experiments, such as non-approved treatment or gene knockout, and then often substituted by mouse. Thus predicting enrichment significance (on a phenotype) of a given gene set of a species (target, say human), by using gene expression measured under the same phenotype of the other species (source, say mouse) is a vital and challenging problem, which we call CROSS-species Gene Set Enrichment Problem (XGSEP). For XGSEP, we propose XGSEA (Cross-species Gene Set Enrichment Analysis), with three steps of: 1) running GSEA for a source species to obtain enrichment scores and p-values of source gene sets; 2) representing the relation between source and target gene sets by domain adaptation; and 3) using regression to predict p-values of target gene sets, based on the representation in 2). We extensively validated XGSEA by using four real data sets under various settings, proving that XGSEA significantly outperformed three baseline methods. A case study of identifying important human pathways for T cell dysfunction and reprogramming from mouse ATAC-Seq data further confirmed the reliability of XGSEA. Source code is available through https://github.com/LiminLi-xjtu/XGSEA Author summaryGene set enrichment analysis (GSEA) is a powerful tool in the gene sets differential analysis given a ranked gene list. GSEA requires complete data, gene expression with phenotype labels. However, gene expression could not be measured under specific conditions for human, due to high risk of direct experiments, such as non-approved treatment or gene knockout, and then often substituted by mouse. Thus no availability of gene expression leads to more challenging problem, CROSS-species Gene Set Enrichment Problem (XGSEP), in which enrichment significance (on a phenotype) of a given gene set of a species (target, say human) is predicted by using gene expression measured under the same phenotype of the other species (source, say mouse). In this work, we propose XGSEA (Cross-species Gene Set Enrichment Analysis) for XGSEP, with three steps of: 1) GSEA; 2) domain adaptation; and 3) regression. The results of four real data sets and a case study indicate that XGSEA significantly outperformed three baseline methods and confirmed the reliability of XGSEA.

bioinformatics↗

bioRxiv2020

BERTMeSH: Deep Contextual Representation Learning for Large-scale High-performance MeSH Indexing with Full Text

Motivation With the rapid increase of biomedical articles, large-scale automatic Medical Subject Headings (MeSH) indexing has become increasingly important. FullMeSH, the only method for large-scale MeSH indexing with full text, suffers from three major drawbacks: FullMeSH 1) uses Learning To Rank (LTR), which is time-consuming, 2) can capture some pre-defined sections only in full text, and 3) ignores the whole MEDLINE database.Results We propose a computationally lighter, full-text and deep learning based MeSH indexing method, BERTMeSH, which is flexible for section organization in full text. BERTMeSH has two technologies: 1) the state-of-the-art pre-trained deep contextual representation, BERT (Bidirectional Encoder Representations from Transformers), which makes BERTMeSH capture deep semantics of full text. 2) a transfer learning strategy for using both full text in PubMed Central (PMC) and title and abstract (only and no full text) in MEDLINE, to take advantages of both. In our experiments, BERTMeSH was pre-trained with 3 million MEDLINE citations and trained on approximately 1.5 million full text in PMC. BERTMeSH outperformed various cutting edge baselines. For example, for 20K test articles of PMC, BERTMeSH achieved a Micro F-measure of 69.2%, which was 6.3% higher than FullMeSH with the difference being statistically significant. Also prediction of 20K test articles needed 5 minutes by BERTMeSH, while it took more than 10 hours by FullMeSH, proving the computational efficiency of BERTMeSH.Contact zhusf{at}fudan.edu.cnCompeting Interest StatementThe authors have declared no competing interest.View Full Text

bioinformatics↗

bioRxiv2019

Viruses of the eukaryotic plankton are predicted to increase carbon export efficiency in the global sunlit ocean

The biological carbon pump, in which carbon fixed by photosynthesis is exported to the deep ocean through sinking, is a major process in Earths carbon cycle. The proportion of primary production that is exported is termed the carbon export efficiency (CEE). Based on in-lab or regional scale observations, viruses were previously suggested to affect the CEE (i.e., viral "shunt" and "shuttle"). In this study, we tested associations between viral community composition and CEE measured at a global scale. A regression model based on relative abundance of viral marker genes explained 67% of the variation in CEE. Viruses with high importance in the model were predicted to infect ecologically important hosts. These results are consistent with the view that the viral shunt and shuttle functions at a large scale and further imply that viruses likely act in this process in a way dependent on their hosts and ecosystem dynamics.

ecology↗

bioRxiv2018

NetGO: Improving Large-scale Protein Function Prediction with Massive Network Information

Automated function prediction (AFP) of proteins is of great significance in biology. In essence, AFP is a large-scale multi-label classification over pairs of proteins and GO terms. Existing AFP approaches, however, have their limitations on both sides of proteins and GO terms. Using various sequence information and the robust learning to rank (LTR) framework, we have developed GOLabeler, a state-of-the-art approach of CAFA3, which overcomes the limitation of the GO term side, such as imbalanced GO terms. Unfortunately, for the protein side issue, available abundant protein information, except for sequences, have not been effectively used for large-scale AFP in CAFA. We propose NetGO that is able to improve large-scale AFP with massive network information. The novelties of NetGO have threefold in using network information: 1) the powerful LTR framework of NetGO efficiently and effectively integrates both sequence and network information, which can easily make large-scale AFP; 2) NetGO can use whole and massive network information of all species (>2000) in STRING (other than only high confidence links and/or some specific species); and 3) NetGO can still use network information to annotate a protein by homology transfer even if it is not covered in STRING. Under numerous experimental settings, we examined the performance of NetGO, such as general performance comparison, species-specific prediction, and prediction on difficult proteins, by using training and test data separated by time-delayed settings of CAFA. Experimental results have clearly demonstrated that NetGO outperforms GOLabeler, DeepGO, and other compared baseline methods significantly. In addition, several interesting findings from our experiments on NetGO would be useful for future AFP research.

bioinformatics↗

bioRxiv2017

Modelling GxE with historical weather information improves genomic prediction in new environments

Interaction between the genotype and the environment (GxE) has a strong impact on the yield of major crop plants. Although influential, taking GxE explictily into account in plant breeding has remained difficult. Recently GxE has been predicted from environmental and genomic covariates, but existing works have not shown that generalization to new environments and years without access to in-season data is possible and practical applicability remains unclear. Using data from a Barley breeding program in Finland, we construct an in-silico experiment to study the viability of GxE prediction under practical constraints. We show that the response to the environment of a new generation of untested Barley cultivars can be predicted in new locations and years using genomic data, machine learning and historical weather observations for the new locations. Our results highlight the need for models of GxE: non-linear effects clearly dominate linear ones and the interaction between the soil type and daily rain is identified as the main driver for GxE for Barley in Finland. Our study implies that genomic selection can be used to capture the yield potential in GxE effects for future growth seasons, providing a possible means to achieve yield improvements, needed for feeding the growing population.

bioinformatics↗

bioRxiv2017

GOLabeler: Improving Sequence-based Large-scale Protein Function Prediction by Learning to Rank

Motivation: Gene Ontology (GO) has been widely used to annotate functions of proteins and understand their biological roles. Currently only {inverted exclamation}1% of more than 70 million proteins in UniProtKB have experimental GO annotations, implying the strong necessity of automated function prediction (AFP) of proteins, where AFP is a hard multi-label classification problem due to one protein with a diverse number of GO terms. Most of these proteins have only sequences as input information, indicating the importance of sequence-based AFP (SAFP: sequences are the only input). Furthermore, homology-based SAFP tools are competitive in AFP competitions, while they do not necessarily work well for so-called difficult proteins, which have {inverted exclamation}60% sequence identity to proteins with annotations already. Thus, the vital and challenging problem now is to develop a method for SAFP, particularly for difficult proteins.\n\nMethods: The key of this method is to extract not only homology information but also diverse, deep-rooted information/evidence from sequence inputs and integrate them into a predictor in an efficient and also effective manner. We propose GOLabeler, which integrates five component classifiers, trained from different features, including GO term frequency, sequence alignment, amino acid trigram, domains and motifs, and biophysical properties, etc., in the framework of learning to rank (LTR), a new paradigm of machine learning, especially powerful for multi-label classification.\n\nResults: The empirical results obtained by examining GOLabeler extensively and thoroughly by using large-scale datasets revealed numerous favorable aspects of GOLabeler, including significant performance advantage over state-of-the-art AFP methods.\n\nContact: zhusf@fudan.edu.cn

bioinformatics↗