Define the object before the label
RNA, peptide and protein tasks require different evidence and different meanings of a negative example.
SUPATCHA RESEARCH · COMPUTATIONAL BIOLOGY · 2013–2026
Read biology.
Build evidence.
From RNA structure to protein destinations and experimentally actionable peptides, this collection follows one research discipline: represent the biological question carefully, test the model honestly, and keep the next experiment visible.
จากโครงสร้าง RNA ไปจนถึงตำแหน่งโปรตีนและ peptide ที่พร้อมให้ห้องทดลองพิจารณา งานชุดนี้ยึดวินัยเดียวกัน คือแทนคำถามทางชีววิทยาให้รอบคอบ ทดสอบโมเดลอย่างซื่อตรง และไม่ลืมว่าคำตอบถัดไปอยู่ในการทดลอง
RNA STRUCTUREROBUSTNESSENSEMBLE LEARNINGPROTEIN LOCALIZATIONTHERAPEUTIC PEPTIDESHARD NEGATIVESEXPERIMENTAL TRIAGE
Build diversity that earns its place
Features and learners are combined because they see complementary signals, not because more models automatically mean better science.
Let evaluation limit the claim
Independent tests, family-aware splits, hard negatives and error analysis define how far a result can travel.
PEER-REVIEWED JOURNAL ARTICLES
Journal Articles / บทความวารสาร
Journal of Molecular Graphics and Modelling · 2026
A mechanism-guided framework for prioritizing membrane-interaction anti-Vibrio peptides from peptidomics data
This work closes the loop from sequence prediction toward experimentally actionable prioritization without hiding the mechanism behind a single opaque score.
Open research dossier / เปิดแฟ้มงานวิจัย →
Life · 2025
EnsembleNPPred: Neuropeptide Prediction Using Ensemble Machine Learning and Deep Learning
Multiple feature views and model families are integrated for robust neuropeptide recognition.
Open research dossier / เปิดแฟ้มงานวิจัย →
BioData Mining · 2022
mSRFR: a machine learning model using microalgal signature features for ncRNA classification
The study moves from generic prediction toward lineage-aware bioinformatics and shows how feature importance can produce a biological hypothesis, not only a score.
Open research dossier / เปิดแฟ้มงานวิจัย →
Frontiers in Genetics · 2022
Ensemble-AHTPpred: A Robust Ensemble Model with a Composite Feature for Antihypertensive Peptides
Composite physicochemical evidence helps identify food-derived peptides that deserve antihypertensive testing.
Open research dossier / เปิดแฟ้มงานวิจัย →
Life · 2021
Ensemble of Multiple Classifiers for Multilabel Classification of Plant Protein Subcellular Localization
A multilabel ensemble respects the possibility that one plant protein can occupy more than one cellular address.
Open research dossier / เปิดแฟ้มงานวิจัย →
Genes · 2021
Ensemble-AMPPred: Robust AMP Prediction Using Ensemble Learning and a New Hybrid Feature
Hybrid biological features and ensemble learning prioritize antimicrobial-peptide candidates while controlling false positives.
Open research dossier / เปิดแฟ้มงานวิจัย →
BioMed Research International · 2019
PSO-LocBact: A Consensus Method for Optimizing Multiple Classifier Results for Bacterial Protein Localization
Particle swarm optimization learns how strongly multiple bacterial localization predictors should be trusted.
Open research dossier / เปิดแฟ้มงานวิจัย →
Nucleic Acids Research · 2014
Identification of Non-Coding RNAs with a New Composite Feature in the Hybrid Random Forest Ensemble Algorithm
A composite SCORE links structure, sequence, modularity, robustness and coding potential across diverse ncRNAs.
Open research dossier / เปิดแฟ้มงานวิจัย →
Nucleic Acids Research · 2013
Heterogeneous Ensemble Approach with Discriminative Features and Modified-SMOTE Bagging for pre-miRNA Classification
Structural robustness, class imbalance and heterogeneous learners meet in a cross-species pre-miRNA detector.
Open research dossier / เปิดแฟ้มงานวิจัย →CONFERENCE PROCEEDINGS
Conference Papers / บทความประชุมวิชาการ
IEEE ICSEC 2025 · 2025
Neuropeptide Classification at Scale with Protein Language Models: Hard Negatives and Cluster-Aware Splits
The contribution is protocol discipline: family-clean evaluation, realistic negatives, transparent baselines and metrics suited to downstream triage.
Open research dossier / เปิดแฟ้มงานวิจัย →
IEEE ECTI-CON 2021 · 2021
Weighted Ensemble for Plant Protein Subcellular Localization Using Particle Swarm Optimization
The paper connects ensemble methodology with deployable workflow design and treats multilabel evidence as information rather than inconvenience.
Open research dossier / เปิดแฟ้มงานวิจัย →
IEEE ICSEC 2019 · 2019
Application of Random Forest in Limited Size Human Long Non-coding RNAs Identification with Secondary Structure Features
The study demonstrates scope discipline: a smaller, biologically coherent question can be more useful than an over-broad classifier trained on limited evidence.
Open research dossier / เปิดแฟ้มงานวิจัย →
ACM ICBEB 2017 · 2017
Identification of Plant Precursor miRNAs using Structural Robustness and Secondary Structures Features
The paper turns RNA folding stability into a measurable feature and shows students how a biological mechanism can guide feature engineering.
Open research dossier / เปิดแฟ้มงานวิจัย →
IEEE ICSEC 2017 · 2017
Improved Prediction of Eukaryotic Protein Subcellular Localization Using Particle Swarm Optimization of Multiple Classifiers
This work establishes the consensus-learning line that later develops into bacterial and multilabel plant localization research.
Open research dossier / เปิดแฟ้มงานวิจัย →
IEEE ICSEC 2017 · 2017
AdaBoost Algorithm with Random Forests for Plant and Animal Precursor MicroRNAs Classification
This paper is a compact lesson in asking a comparative biological question rather than treating every pre-miRNA as one homogeneous class.
Open research dossier / เปิดแฟ้มงานวิจัย →
IEEE ISMS 2016 · 2016
Enhanced Viral Precursor MicroRNA Identification with Structural Robustness Features in Back-propagation Neural Network
The study is an early example of using perturbation-sensitive structural evidence when alignment-based signals become unreliable.
Open research dossier / เปิดแฟ้มงานวิจัย →