Li Chen

Dr. Li Chen


Research Focus & Metrics

Academic Career Length

1Year (2026-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
20 June 2026 in Life
Verified
Journal Article
Open access
PDF file
MAMia Yang Ang
Mia Yang Ang
LCLi Chen
Li Chen
LSLanni Song
Lanni Song
LEONARD LIPOVICHLEONARD LIPOVICH
LEONARD LIPOVICH
SCSiew Woh Choo
Siew Woh Choo
Large language models (LLMs) are increasingly adopted in life-science research for scientific writing, coding, literature synthesis, workflow troubleshooting, and preliminary data interpretation. In microbial genomics and bioinformatics, their appeal is clear because researchers routinely integrate genome annotations, antimicrobial resistance profiles, virulence determinants, taxonomic assignments, microbiome outputs, workflow scripts, and primary literature. Yet this domain also highlights major risks, including hallucinated biological claims, inaccurate citations, irreproducible code, unsupported genotype-to-phenotype inference, and inappropriate clinical or public health framing. This narrative review examines responsible LLM use in microbial genomics as a representative life-science setting where interpretation depends on database provenance, validated workflows, expert assessment, and reproducible evidence chains. It considers applications in genome annotation, antimicrobial resistance interpretation, virulence analysis, microbiome and metagenomics workflows, coding support, and scientific writing. The review further presents MicrobeGuardGPT as a conceptual reliability framework for assessing LLM-assisted microbial genomics outputs before scientific, clinical, or public health use. By connecting task domains, evidence verification, expert validation, and reliability classification, the framework supports risk-aware LLM integration in bioinformatics. Responsible implementation will require domain-specific benchmarks, curated database linkage, transparent reporting, reproducible workflows, human oversight, and governance standards tailored to biological interpretation across research, diagnostic, surveillance, outbreak-response, educational, and translational contexts.

14 July 2026 in Biomolecules
Verified
Journal Article
Open access
PDF file
MAMia Yang Ang
Mia Yang Ang
LCLi Chen
Li Chen
LSLanni Song
Lanni Song
LEONARD LIPOVICHLEONARD LIPOVICH
LEONARD LIPOVICH
SCSiew Woh Choo
Siew Woh Choo
Genome-wide association studies (GWAS) have identified thousands of loci associated with complex human diseases. However, the majority of the association signals reside in non-coding regions of the genome, and do not directly reveal the causal variant, effector gene, regulatory biomolecule, cell type, pathway, biomarker, or therapeutic target. Because many disease-associated variants act through non-coding regulatory mechanisms, post-GWAS interpretation increasingly depends on fine-mapping, expression quantitative trait loci, transcriptome-wide association studies, and functional evidence from single-cell multi-omics, network biology, and genetic target prioritization. Artificial intelligence can attempt to integrate these heterogeneous molecular evidence layers, but the resulting black-box prediction is insufficient when outputs cannot be biologically reproduced or experimentally tested. This review evaluates explainable artificial intelligence (XAI) as a framework for linking genetic association signals to molecular mechanisms and causal gene hypotheses. We argue that explainability is best treated as a biological requirement because useful models must expose evidence paths from significant disease-associated variants to regulatory elements, genes, transcripts, proteins, pathways, cell states, and therapeutic hypotheses. By emphasizing transparent evidence provenance, ancestry-aware interpretation, and functional validation, XAI can support the translation of GWAS signals into molecularly testable hypotheses for target prioritization and precision molecular medicine. The review focuses on the question of how to accomplish AI-accelerated functionalization of GWAS outputs across complex human diseases and traits.

4 September 2026 in Biology
Verified
Journal Article
Open access
PDF file
MAMia Yang Ang
Mia Yang Ang
LEONARD LIPOVICHLEONARD LIPOVICH
LEONARD LIPOVICH
SCSiew Woh Choo
Siew Woh Choo
LCLi Chen
Li Chen
LSLanni Song
Lanni Song
Agentic artificial intelligence is extending bioinformatics beyond conversational assistance by enabling systems to select tools, execute code, revise analytical plans, and interpret biological data. These capabilities may accelerate research, but they also redistribute decisions that determine whether biological conclusions are valid. We conducted a targeted, structured PubMed search in July 2026 and identified 11 peer-reviewed agentic bioinformatics systems for descriptive review based on predefined eligibility criteria for analytical decision-making, tool or code execution, iterative evaluation, or coordinated agent activity. The evidence base covered single-cell transcriptomics, microbial genomics, cancer genomics, and omics applications, together with methodological literature on reproducibility and biological validation. We examined how current systems report delegated authority, provenance, validation, evidence, abstention, and human oversight. Existing platforms implement safeguards such as sandboxed execution, restricted commands, interaction logs, evidence identifiers, automated checks, critic agents, quality scores, and expert assessment. However, published reports rarely provide a connected account linking the original biological question to samples, reference resources, analytical decisions, computational actions, statistical results, supporting evidence, validation outcomes, and final claims. We distinguish inherited bioinformatics errors, errors amplified through autonomous action, and emergent failures arising from memory, retrieval, tool interaction, or agent coordination. We further propose a multidimensional decision-rights profile, consequence-sensitive validation gates, and a claim-to-evidence provenance architecture organized through the Traceable History of Research Evidence, Agent Actions, and Decisions in Bioinformatics (THREAD-Bio) framework. Illustrative cases show that technically successful execution may still support misleading inference. Trustworthy agentic bioinformatics therefore requires claims to remain reconstructible, challengeable, validated, and proportionate to the evidence.

16 September 2026 in Pharmaceuticals
Verified
Journal Article
Open access
PDF file
MAMia Yang Ang
Mia Yang Ang
LCLi Chen
Li Chen
LSLanni Song
Lanni Song
LEONARD LIPOVICHLEONARD LIPOVICH
LEONARD LIPOVICH
SCSiew Woh Choo
Siew Woh Choo
Background/Objectives: Antimicrobial resistance has renewed interest in live therapeutic phages, including engineered candidates and defined phage cocktails, as antibacterial strategies. However, clinical translation requires evidence extending beyond phage isolation and computational prediction. This review examines how phage genomics and bioinformatics can support phage-based antibacterial development while remaining aligned with pharmaceutical requirements for safety, quality, pharmacology, and clinical validation. Methods: This narrative review synthesizes the literature on the generation, interpretation, and experimental validation of genomic and bioinformatic evidence for live therapeutic phages. Representative approaches for viral identification, genome-quality assessment, annotation, comparative genomics, lytic–temperate lifestyle classification, host prediction, receptor analysis, antiphage-defence profiling, and artificial-intelligence-assisted prioritization were evaluated according to their outputs, principal failure modes, validation requirements, and supported development decisions. Results: Current tools address distinct analytical tasks. Examples include VIBRANT and geNomad for viral identification, CheckV and PhageTerm for genome-quality and termini assessment, Pharokka, PHANOTATE, and PHROGs for gene prediction and annotation, PhageAI, BACPHLIP, and PhaTYP for lytic–temperate lifestyle prediction, iPHoP and CRISPR spacer matching for host prioritization, and PADLOC and DefenseFinder for bacterial defence profiling. Their outputs differ in taxonomic resolution, reference coverage, training-data dependence, and biological interpretation. The review maps these outputs to proportionate validation requirements while integrating formulation and PK/PD context, sequence-to-product traceability, intellectual-property documentation, minimum-information reporting, and resistance-responsive redesign. Conclusions: This review presents a practical sequence-to-product framework for interpreting contemporary phage-bioinformatics methods. By separating computational discovery from isolate-level activity, product quality, pharmacological evidence, and clinical monitoring, it clarifies the decisions supported at each stage and the additional evidence required before candidate progression, product use, or redesign.

2 April 2026 in bioRxiv
Verified
Preprint
Open access
PDF file
ZHZhizhou He
Zhizhou He
YLYibai Li
Yibai Li
TSTatiana P. Shkurat
Tatiana P. Shkurat
EBElena V. Butenko
Elena V. Butenko
EDEkaterina G. Derevyanchuk
Ekaterina G. Derevyanchuk
LLLeonard Lipovich
Leonard Lipovich
Background: Polycystic ovary syndrome (PCOS) is a prevalent endocrine disorder and a leading cause of female infertility, with complex genetic, metabolic, and hormonal etiologies. Long non-coding RNAs (lncRNAs) have emerged as important regulators of diverse biological processes, yet their roles in PCOS remain underexplored. Here, we identified and characterized PCOS differentially expressed gene-associated lncRNAs (PDEGAL) with an integrative approach combining expression data, genetic association, and evolutionary analysis. Methods: Thirty-three PCOS-associated protein-coding genes were obtained from our prior study, and all their nearby and overlapping lncRNAs were annotated. These candidates were analyzed using UCSC Genome Browser-mapped annotations and datasets, including NCBI RefSeq, GENCODE, GTEx, GWAS SNPs, and conservation, as well as the FANTOM5 cap analysis of gene expression (CAGE) promoter data, to assess their expression, regulatory potential, genetic variant overlaps, and evolutionary conservation. Results: Twenty-three PDEGALs (18 antisense to, and 5 sharing bidirectional promoters with, known PCOS-associated protein-coding genes) were identified. 17 PDEGALs contained GWAS SNPs with statistically significant disease associations, 9 of which were associated with PCOS-related traits. 5 PDEGALs demonstrated expression in the KGN granulosa cell model of PCOS. Key gene structure element (KGSE) analysis revealed that most PDEGALs are primate-specific. Integrating four criteria—GTEx expression, GWAS SNPs, FANTOM promoterome, and KGSE conservation—highlighted HELLPAR as the only lncRNA fulfilling all four, while five others—PGR-AS1, MTOR-AS1, ENSG00000265179, ENSG00000256218, and LOC105377276—fulfilled three of the four criteria. Conclusions: We have systematically identified candidate PCOS regulatory lncRNAs with convergent genetic, expression, and evolutionary evidence. These results provide a framework for functional validation and highlight lncRNAs as potential biomarkers and therapeutic targets in PCOS that function by regulating their nearby and overlapping protein-coding genes.

12 May 2026 in Biology
Verified
Journal Article
Open access
PDF file
MAMia Yang Ang
Mia Yang Ang
LCLi Chen
Li Chen
HPHeru Pramono
Heru Pramono
TLTeck Yew Low
Teck Yew Low
Nur Azalina Suzianti FeisalNur Azalina Suzianti Feisal
Nur Azalina Suzianti Feisal
SCSiew Woh Choo
Siew Woh Choo
Foodborne diseases present a serious public health challenge, causing roughly 600 million illnesses and 420,000 deaths annually. A significant portion of this impact is felt in Asia, where traditional fermented and dry-salted seafood, such as katsuobushi, budu, and peda, are dietary staples. These products rely on diverse microbial communities that determine their final safety, flavor, texture, and shelf life. Historically, research has centered on lactic acid bacteria (LAB), yet the functional contributions of non-LAB halotolerant species, including genera like Tetragenococcus, Staphylococcus, and Bacillus, are functionally important in these high-salinity niches. This review evaluates the transition from basic taxonomic surveys to mechanistic multi-omics approaches, integrating genomics, transcriptomics, proteomics, and metabolomics to decode microbial functionality under selective environmental pressures. We discuss how genomic mining using platforms such as BAGEL4 and antiSMASH can uncover biosynthetic gene clusters and antimicrobial peptides, while CARD supports antimicrobial resistance monitoring. Transcriptomic analysis reveals microbial responses to osmotic stress, low water activity, and pH fluctuations, whereas proteomic profiling links gene expression to active enzymes, stress proteins, and functional biomarkers. Metabolomics captures the chemical outcomes of fermentation, including amino acids, volatile organic compounds, spoilage markers, and biogenic amines. By merging these high-dimensional datasets with artificial intelligence, researchers can move toward predictive modeling that distinguishes biological causation from simple correlation. This shift offers a strategy to improve the safety, consistency, and resilience of traditional high-salinity fermented seafood systems.