Note: PST2026

https://pst2026.md.chula.ac.th/#

Proteome - Nanopore focus

-- 5% in nucleus is transcription factors which related to the disease. Currently, ~ 10 drugs targeting transcription factor.

-- TF that not response well to the drug -- 1. majorly appears as complex proteins 2. low abundance (มีปริมาณน้อยมาก) 3.highly disordered ถ้าจะ predict ด้วย alphafold ก็จะทำได้ยาก เพราะว่าโครงสร้างไม่ solid look like a spagetti - hardly to form 3D (I guess it requires substrate to bind before forming the solid 3D)

-- แลบนี้มีความพยายามที่จะหายาที่ target TF เป็นบริษัท starup ที่ทำเกือบทุกอย่างเป็นระบบ automate

--------

Nanopore

-- เหมาะสำหรับการทำ Single cell

-- ตัว pore มีอยู่ 2 ประเภท 1.biological type 2. solid state nanopore (หลัก ๆ งานจะใช้ตัวนี้เนื่องจากจะใช้กับงาน protein sequencing)

-- ในการอ่านลำดับโปรตีน -- ต้องทำให้โครงสร้างมันเป็สายเดี่ยว ๆ ก่อนก่อนที่จะลง pore ซึ่งเทคโนโลยีตอนนี้จะมี enzyme ที่ช่วยในการคลายโครงสร้างดังกล่าวอยู่แล้ว

-------

KRAS -- in CCA

-- Mutation signature ของ case เมืองนอก กับเมืองไทยแตกต่างกัน ดังนั้นถ้าจะใช้ precision medicineยาที่ใช้กับกลุ่ม CCA เมืองนอก กับเมืองไทย มีความแตกต่างกันเนื่องจาก driver คนละตัว

-- สำหรับกลุ่ม KRAS -- เป็นโปรตีนที่มีหลายหน้าที่ ยังไม่เป็นที่รู้จักมากนักว่าทำอะไรใน CCA แน่ ๆ เกี่ยวข้องกับ cell cycle และการควบความปริมาณโปรตีนภายในเซลล์ โดย favor ไปในทางการผลิตโปรตีน

-- G12C -- มีประมาณ 10% ที่ mutate ใน CCA แต่จริง ๆ แล้ว มีตัวอื่น ๆ อีก แต่ไม่ได้จดมา

------

Cryo-EM

-- แนะนำให้ทำกับตัวใหญ่เพราะหลักการของมันจะใช้ในเรื่องการ contrast ของแสง ดังนั้นที่เขาแนะนำคือ 100 kDa จะทำให้ resolve structure ได้ดีกว่า

-- TEM with Cryo-EM -- ใช้สำหรับการดูเรื่อง quality ก่อนที่จะนำไปทำ Cryo-EM

-- โปรตีนใช้น้อยมากระดับ ng แต่ตัวอย่างต้อง purify ค่อนข้างดี (homogenous sample)

-- Shotgun | bottom up Cryo-EM pipeline -- > เอามาทำพร้อม ๆ กันกับ MassSpec กระบวนการคือ break cells ก่อน แล้วทำ partial seperation ก่อนที่จะเอาไปเข้า MassSpec and Cryo-EM แบบ pararelle แล้วเอามา match กัน

-----

From NotebookLM

Based on a detailed extraction of the presentations and records from the 20th International Symposium of the Protein Society of Thailand (PST2026), here is the comprehensive, scientific breakdown of each speaker's research, clinical trials, methodology, and key results.

1. Associate Professor Michal Bassani-Sternberg

University of Lausanne (UNIL) & Ludwig Institute for Cancer Research, Lausanne, Switzerland

Presentation Focus: From Discovery to Therapy: Overcoming Challenges in Neoantigen-Driven Cancer Immunotherapy

The NeoDisc Proteogenomic Pipeline

The speaker's group developed NeoDisc, a rapid, modular, end-to-end clinical proteogenomic pipeline designed to prioritize tumor-specific antigens. Unlike older computational tools that function strictly in isolation, NeoDisc directly integrates whole-exome sequencing (WES), transcriptomics (RNA-seq), and mass spectrometry-based immunopeptidomics with advanced in silico tools to identify canonical neoantigens, viral antigens, and non-canonical tumor-specific peptides.

Timelines: The computational pipeline identifies and ranks the candidate peptides, producing a prioritized target list within approximately 2 weeks of receiving patient biopsies.
Clinical Production: Synthesizing the peptides takes about 3 weeks, making the entire workflow from biopsy to final clinical vaccination/infusion ready in 5 to 6 weeks.

Phase I Clinical Trial Insights

In collaboration with Professor George Coukos, they conducted a Phase I clinical trial using personalized adoptive T-cell therapies. Between 50 and 100 targets were identified per patient, synthetically manufactured, and used to expand tumor-infiltrating lymphocytes (TILs) into billions of cells for re-infusion. Through this trial, they mapped three distinct clinical phenotypes of treatment response and resistance:

Antigen Presentation Defects (Disease Progression): In a patient who experienced disease progression one year post-treatment, phylogenetic mapping of longitudinal biopsies revealed a loss of heterozygosity (LOH) in beta-2 microglobulin ((B2M)). Losing one copy of the (B2M) gene, paired with secondary mutations, completely abolished HLA class I presentation on the tumor surface, allowing cancer cells to escape T-cell recognition.
Gene Downregulation (Mixed Clinical Response): In a lung cancer patient with highly mixed lesion responses, 60% of the infused CD8+ T-cells targeted a single, highly immunogenic mutation. While some lesions disappeared, progressing lesions escaped immunological pressure by selectively downregulating the transcript expression of the specific gene harboring that mutation.
Complete Clinical Cure: Their first treated melanoma patient achieved a complete response, with no detectable cancer cells in biopsies within 30 days. Although they identified a highly dominant T-cell receptor (TCR) clone driving this cure, its cognate target was mapped to a non-canonical open reading frame, pointing to shared, unannotated tumor antigens as valuable therapeutic targets.

Targeting the Non-Canonical Immunopeptidome

To expand the antigen pool beyond rare somatic mutations, the lab is targeting transposable elements (TEs). These elements are typically silenced in healthy tissue but aberrantly active in cancer. They act as unique cancer-specific promoters or produce alternatively spliced junction peptides with coding potential. By performing deep sequencing on a cohort of 67 patients, they discovered 45 tumor-exclusive transposable element-derived peptides. They capture these non-canonical targets by integrating ribosome profiling (Ribo-seq) to identify novel active translation products and searching translated three-frame open-reading-frame databases.

2. Dr. Kasidet Manakongtreecheep

Broad Institute of MIT and Harvard, Dana-Farber Cancer Institute, and Harvard Medical School

Presentation Focus: Sensitive Direct Detection of Cancer Antigens Enabled by User-Defined Peptide Libraries

Pepyrus (Genetic Peptide Library Synthesis)

Traditional mass spectrometry (specifically Data-Dependent Acquisition, or DDA) struggles with low sensitivity and a high rate of peptide "dropouts" when scanning for rare, lowly abundant neoantigens. To resolve this, the speaker developed Pepyrus, an E. coli-based genetic engineering platform capable of synthesizing user-defined, patient-specific peptide libraries at low cost and massive scale.

DIA-MS Integration: They synthesized libraries of over 100,000 patient-specific and shared cancer peptides. By acquiring high-resolution reference spectra of these physical libraries, they built clean, tailored spectral databases. These are used to deconvolute complex, co-fragmented spectra generated by Data Independent Acquisition mass spectrometry (DIA-MS).
Detection & Recycling: This approach successfully rescued rare neoantigens, unannotated open-reading frames, and endogenous retroviral peptides from clinical samples that were completely missed by conventional DDA. Because reference spectra reside permanently in a digital database, these libraries can be recycled "off-the-shelf" to deconvolute mass spectra across different patients, even with low cell inputs down to 1 million cells.

HLA-flex (Physical Binding Mapping)

To map peptide-HLA binding without relying on flawed prediction algorithms, they developed HLA-flex. This platform co-expresses user-defined peptide libraries with recombinant HLA molecules and positive folding chaperones inside E. coli. They successfully constructed a pilot library of 10,000 peptides across 10 prevalent class I HLA alleles. HLA-flex demonstrated >90% detection accuracy of validated binders, while also identifying physical binders that algorithms predicted would not bind, highlighting the limitations of in silico-only models.

E. coli Utility: Using E. coli rather than mammalian cells allows high scalability (using 500 mL to 1L bioreactors) to yield high concentrations of physical peptides. This allows researchers to split a single batch to run parallel assays under different conditions, such as testing thermal stability across multiple heat treatments.

Translational Focus: HLA-E and DLBCL

Unlike classical HLA class I alleles, non-classical HLA-E has only two major alleles (E*01:01 and E*01:03) and primarily presents signal peptides to natural killer (NK) cells to inhibit cytotoxicity. However, HLA-E can also act as an activating receptor under specific disease contexts. Targeting Diffuse Large B-cell Lymphoma (DLBCL)—which exhibits extremely high HLA-E expression—the lab performed HLA-E immunopurification on primary patient samples and PDX models. They built a DLBCL-specific database, co-expressed candidate peptides using the HLA-flex system, and identified novel tumor-associated HLA-E-bound peptides derived from key oncogenes, including BCL11A and BCL6.

3. Senior Researcher Dr. Poorichaya Somparn

Center of Excellence in Systems Biology, Faculty of Medicine, Chulalongkorn University, Thailand

Presentation Focus: Mass Spectrometry-Based Profiling of Personalized Immunopeptidomes in Thai Renal Cell Carcinoma

Addressing the Population HLA Gap

Most public immunopeptidome databases are constructed using data from Western populations, resulting in a severe representation bias. Southeast Asian populations exhibit highly distinct HLA class I distributions. For example, in their Thai cohort, HLA-A*11:01 was the most frequent allele, occurring in 69% of patients. They identified seven alleles in their cohort that had never been documented in major public immunopeptidome databases.

Study Methodology & Somatic Neoantigen Discovery

The study profiled the personalized immunopeptidomes of 13 Thai patients with renal cell carcinoma (RCC). They utilized a proteogenomic approach: matched tumor and normal tissue whole-exome sequencing (WES) was used to construct patient-specific reference databases, followed by HLA class I immunoaffinity purification and LC-MS/MS analysis.

The High Hurdle of Neoantigen Isolation: Identifying physical somatic neoantigens in RCC remains a massive technical challenge. Across the entire 13-patient cohort, database-driven searches identified only a single somatic mutated neoantigen.
The Mutant JADE2 Peptide: This single neoantigen was isolated from patient RCC5, who possessed the highest baseline tumor mutational burden (TMB) in the cohort. The neoantigen is a 9-mer peptide mapping to a mutated JADE2 protein. The somatic mutation caused a threonine-to-serine substitution at protein position 391, which was computationally predicted to bind strongly to the patient-specific HLA-B*40:01 allele.

Validation and Immunogenicity

To rule out false identification, they synthetically manufactured the mutant JADE2 peptide. The synthetic peptide showed a perfect match with the patient's MS/MS spectrum and eluted at the exact same 30-minute retention time. Next, they tested its functional immunogenicity by co-culturing the peptide with patient-derived dendritic cells and PBMCs. The mutated JADE2 peptide induced a highly significant increase in interferon-gamma secretion in patient T-cells compared to its wild-type counterpart. Importantly, no response was induced in HLA-matched healthy controls, confirming that the patient had established tumor-specific immunological priming.

4. Dr. Khin Zay Yar Myint

Centre for Advanced Medical Science and Technology, Tokyo Midtown Medical Center, Japan

Presentation Focus: Factors Associated with the Survival Rates of Patients Receiving a Dendritic Cell Vaccine Pulsed with Different Cancer-Associated Antigens: A Retrospective Observational Study

Study Cohort and Design

This real-world, retrospective study evaluated 124 cancer patients treated at their center between March 2020 and February 2024. The cohort comprised 63 men and 61 women aged 34 to 91 years. Cancers treated included pancreatic cancer (the majority), followed by breast, colon, and stomach cancers. Patients were treated with dendritic cell (DC) vaccines (typically 7 doses) pulsed with one of two antigen types:

WT1 Peptide (n = 66): A widely-expressed cancer antigen targeting WT1 (often co-pulsed with MUC1 or CEA in adenocarcinomas), which acts as a common cancer stem cell target.
Personalized Neoantigens (n = 58): Custom-designed peptides matched to the patient’s somatic mutations.

Liquid Biopsy Monitoring

The team utilized two commercial liquid biopsy assays—Guardant360 and Genotype (referred to as General Dive)—to profile circulating tumor DNA (ctDNA). These tests were administered at baseline (pre-vaccination) and at follow-up (after the 4th DC dose). Guardant360 was preferentially selected for patients with a short life expectancy (<6 months) due to its faster turnaround time.

Clinical Findings & Prognostic Markers

Safety: Both vaccine formulations were extremely safe, with adverse events limited to transient, mild Grade 1 or 2 fevers or local injection-site reactions.
Prognostic Predictors (Multivariate Analysis): Survival outcomes were determined by baseline clinical/genomic characteristics rather than the type of vaccine antigen used:
- Favorable Prognosis: Associated with female sex, a baseline ctDNA ratio < 1%, early-stage cancers, and a treatment intent of prevention of recurrence (post-resection). Notably, patients who received neoantigen-pulsed DC vaccines for recurrence prevention achieved a 100% survival rate over the follow-up period.
- Poor Prognosis: Associated with male sex, advanced therapeutic intent, older age, high baseline ctDNA, an increasing ctDNA ratio at follow-up, and the emergence of new somatic mutations. Patients with "Group 1" cancers (pancreatic, lung, biliary tract, gallbladder) showed high hazard ratios and poor outcomes.

WT1 vs. Neoantigen Clinical Trade-Offs

While personalized neoantigen vaccines provide highly targeted therapy, their manufacturing workflow from liquid biopsy results takes 45 days. Conversely, the WT1 peptide is pre-fabricated and immediately available, allowing vaccination to begin within 2 to 4 weeks. Consequently, Tokyo Midtown Medical Center doctors strongly recommend WT1-pulsed vaccines for patients with advanced, highly aggressive disease or short life expectancies, while reserving personalized neoantigen vaccines for patients in stable or preventative clinical phases.

-----------------

Academic Presentation: Asst. Prof. Sira Sriswasdi

Center of Excellence in Computational Molecular Biology, Faculty of Medicine, Chulalongkorn University, Thailand

Topic: Improving peptide-HLA immunogenicity prediction with protein structure-informed deep learning
Speaker Profile: Asst. Prof. Sira Sriswasdi (referred to by colleagues as John Sira, Suri, or Sula) completed his undergraduate studies at MIT, his PhD at the University of Pennsylvania and the Wistar Institute, and his post-doctoral training at the University of Tokyo. He serves as the founding director of the Center of Excellence for Computational Molecular Biology and is structurally characterized as a mathematician who applies advanced computational frameworks to address biological and biochemical problems.

1. Error Correction in De Novo Peptide Sequencing

The De Novo Sequencing Bottleneck: Liquid chromatography-tandem mass spectrometry (LC-MS/MS) based de novo peptide sequencing is highly prone to errors. For a standard 12-mer peptide, it is common for 2 or 3 amino acids to be incorrectly identified, which presents a massive bottleneck for clinical neoantigen vaccines where sequence precision is critical.
Mass-Based Decoupling Pipeline: To eliminate sequencing hallucinations and prevent the model from generating incorrect amino acids, Sira's lab developed a pipeline where low-confidence amino acid calls are converted into their corresponding exact masses. If the user possesses a reference database, these combined masses are then decoded back to find the true sequence, resulting in a high-confidence recovery of de novo peptides.
Shifting Decision Levels: To further ensure sequence quality, candidate peptides are validated against reference databases (such as the human proteome), shifting the classification decision from the individual peptide level to the protein level.
Non-Clinical Applications: Beyond oncology, this pipeline is deployed in collaborative projects to identify novel peptides in indigenous Thai species and fish.

2. Uncovering HLA Binding Motifs and Sub-Motifs

Clustering Revealed Motifs: Standard immunopeptidome databases simplify ligand motifs into a single consensus pattern per HLA allele. By performing rigorous, multi-level clustering on MS-identified peptides, Sira’s group discovered that individual alleles present multiple, structurally distinct binding sub-motifs.
Allele Examples:
- HLA-B*40:02 (referred to as B4 O2 / 140): Displays sub-motifs with distinct anchor residues, such as possessing an acidic residue (aspartate/glutamate) at position 2, or harboring the acidic residue in the middle of the peptide.
- HLA-B*51:01 (referred to as HP 5101): Reveals clear sub-populations of ligands where some motifs present aspartic acid at position 1, whereas other sub-motifs completely lack it.

3. Overcoming Western Population Bias in HLA Prediction

Addressing the Population Gap: Most public peptide-MHC binding prediction tools are trained on datasets heavily dominated by Western HLA distributions, rendering them inaccurate for Southeast Asian alleles. Sira's machine learning model was trained on comprehensive datasets spanning both Thai and non-Thai populations.
Outperformance on Rare Alleles: This approach achieved drastic improvements in prediction performance for rare, underrepresented HLA alleles that have highly limited training data in public databases.

4. Structure-Informed Immunogenicity Prediction

Predicting T-cell Activation: Sira noted that predicting whether a presented peptide is actually "immunogenic" (capable of activating T-cells to kill tumor cells) is a significantly harder problem than predicting HLA binding.
Predicting Without TCR Sequencing: To make prediction clinically viable without requiring extremely complex sequencing of a patient's individual T-cell receptor (TCR) pool, Sira's lab assumed each patient possesses a broad, representative TCR pool.
AlphaFold2 and Molecular Dynamics (MD): They trained a deep learning classifier using only the structures of the HLA molecule and the peptide. At the end of the pandemic, they pioneered this by modeling peptide-MHC structures from scratch using AlphaFold2 combined with computationally heavy Molecular Dynamics (MD) simulations. (Sira noted that a similar structure-informed deep learning concept was published by another research group in Nature Machine Intelligence earlier this year).
Explainable AI (XAI): To understand the biophysical rules driving the model, they perform in silico mutagenesis, substituting individual residues and tracking whether the model’s prediction confidence drops. If changing a specific residue causes the prediction confidence to plummet, it is identified as a critical immunogenic hotspot. Sira confirmed they currently rely on AlphaFold2 for these explainability pipelines but plan to transition to AlphaFold3 once their current studies are finalized.

5. Technical Q&A Highlights

Integrating RNA Expression: Sira confirmed that they integrate RNA expression data with genomic mutations to ensure that any prioritized mutant neoantigen is actively transcribed and expressed by the tumor cells.
WGS vs. WES: Whole-Genome Sequencing (WGS) is currently too expensive to be performed routinely on all clinical tumor biopsies, so they prioritize matched Whole-Exome Sequencing (WES) for tumors, though they perform germline variant sequencing on patient blood samples.
Fusion Genes and degradation: They have completed fusion gene analysis on total RNA-seq data. However, because many clinical patient samples are long-term archived tissues, RNA degradation is a major hurdle that can limit detection sensitivity.

Sponsor Presentation: Mr. Haris Irshad

Protein Research Sales Specialist, Biacore (transcribed as Bernard Trading)

Topic: Accelerating discovery with speed, scalability, and sensitivity by the Biacore 8 Series SPR systems

1. Parallel Screening and 16 Flow Cell Utilization

Surface Plasmon Resonance (SPR): The presentation highlighted the capabilities of the newly launched Biacore 8 Series SPR systems designed to optimize antibody drug discovery workflows.
Parallel Flow Channels: While previous generations of SPR systems had routing limitations, the 8 Series allows researchers to utilize all 16 flow cells across 8 parallel channels simultaneously.
High-Throughput Off-Rate Ranking: This allows rapid, head-to-head off-rate screening. For example, antibodies 1 through 8 can be captured on the first flow cell of each of the 8 channels, and antibodies 9 through 16 can be routed to the second flow cell of each channel.

2. Triple-Parameter Single-Run Assay

Multi-Tasking Workflow: The Biacore 8 Series can measure three independent parameters—concentration, kinetics, and epitope binning—within a single, unified run.
Resource and Time Savings: By consolidating these assays, researchers cut their total run time by approximately 50% (reducing a traditional 13-hour workflow down to just 7 hours) while saving half of their precious sample volume.

3. Innovations in Concentration Analysis

The system supports three distinct modes of active concentration measurement depending on sample constraints:

Serial Concentration Analysis: Generates a dedicated calibration curve for each of the 8 channels individually, which is accurate but highly sample- and reagent-intensive.
Parallel Concentration Analysis: Utilizes a single, shared calibration curve across multiple channels while incorporating normalizations and controls to account for channel-to-channel variations, significantly saving reagents.
Rapid Concentration (8 Series Innovation): Focuses on rapid screening by performing ultra-short sample injections lasting only 1 to 2 seconds. It leverages the shared calibration curve model but removes redundant wash steps, leaving nearly the entire 384-well plate open for active screening. This method can fully process 384 samples in just 24 minutes, compared to 4 hours using traditional serial methods.

4. Software & Analysis Extensions

The system runs on the Biacore Insights software platform. It incorporates dedicated analysis extensions to accelerate data processing:

Extended Screening Extension: Optimized for evaluating small molecules, fragments, and antibody leads.
Concentration and Potency Extension: Integrates seamlessly with epitope binning assays to quickly characterize candidate molecules.

-----------------

Voice 003 sd: Comprehensive Scientific Session Report

Session: Parallel Session Mandarin A (Protein Modification and Cell Signaling) & Parallel Session Mandarin B (Synthetic and Chemical Biology) The 20th International Symposium of the Protein Society of Thailand (PST2026)

1. Assistant Professor Naphat Chantaravisoot

Department of Biochemistry, Faculty of Medicine, Chulalongkorn University, Thailand

Presentation Topic: Decoding Tumor Aggressiveness Through Phosphoproteomics: Convergent mTORC2 Signaling in Cancers

Mechanistic and Multi-Omic Framework

The mTORC2 Complex: While the mechanistic target of rapamycin complex 1 (mTORC1) has been extensively characterized in oncology, the complex 2 (mTORC2) remains poorly understood regarding its role in driving the aggressive features of glioblastoma (GBM).
Multi-Omics Methodology: The lab utilized an integrative multi-omics approach—combining transcriptomics, proteomics, phosphoproteomics, and interactomics—to map the downstream signaling cascades of mTORC2 in GBM cells. These analyses identified mTORC2 as a critical signaling hub regulating cell migration, cytoskeletal remodeling, phenotypic plasticity, and genome instability.

Discovery of the mTORC2–AKT–BABAM1 Axis

The DNA Repair Link: Quantitative phosphoproteomics and interactomics identified a direct link between mTORC2 and the cellular DNA damage response. Specifically, they characterized the mTORC2–AKT–BABAM1 (also known as MERIT40) signaling axis.
Biophysical Mechanism: When the integrity of mTORC2 is complete, it activates AKT, which directly phosphorylates BABAM1 (MERIT40). Phosphorylated BABAM1 is recruited into the nucleus where it forms the BRCA1-A regular complex to repair double-stranded DNA breaks.
Knockdown and Mutagenesis: Silencing the crucial mTORC2 component Rictor (RTOR) or performing a non-phosphorylatable S29A mutagenesis of BABAM1 completely blocked BABAM1's nuclear translocation. This prevented the formation of the BRCA1-A repair complex, resulting in a catastrophic failure of double-stranded break repair and significantly increased apoptosis rates.

Synergistic Chemotherapy Reversal

Chemotherapy Resistance: Temozolomide (TMZ) is the standard alkylating chemotherapeutic agent for GBM, but tumors frequently develop resistance by hyperactivating their DNA repair machinery.
Dual Inhibition Synergy: In both 2D and 3D glioblastoma spheroid models (using highly resistant U87-MG cell lines), administering sub-lethal doses of TMZ actually stimulated the phosphorylation of AKT and BABAM1 as a survival mechanism.
Overcoming Resistance: Combining TMZ with an ATP-competitive dual mTORC1/2 inhibitor—specifically vistusertib (AZD2014) or AZD8055—abolished this reactive phosphorylation. This synergistic block prevented BRCA1 recruitment to the nucleus, successfully restricting the growth and volume of the resistant 3D tumor spheroids and triggering apoptosis.

2. Dr. Jutatip Panaampon

Department of Medical Oncology, Dana-Farber Cancer Institute, Harvard Medical School, Boston, USA

Presentation Topic: DUSP2 promotes lymphomagenesis through CDK1 activation and enhanced lymphoid cell proliferation

Dual Role of DUSP2 in Malignancies

Context: Dual-specificity phosphatase 2 (DUSP2, historically called PAC-1) is a nuclear threonine/tyrosine phosphatase whose expression is highly restricted to immune cells. While it has been proposed to act as a tumor suppressor in some solid tumors, its role in mitotic proliferation and hematologic malignancies had not been systematically characterized.
Overexpression in Lymphomas: Bioinformatic and tissue analyses revealed that DUSP2 is highly overexpressed in human B-cell, T-cell, and other hematologic malignancies, especially germinal center B-cell (GCB) diffuse large B-cell lymphoma (DLBCL).

Functional Knockdown and Transgenic In Vivo Models

Ablation: Genetically ablating or knocking down DUSP2 in human lymphoma cell lines (including DHL4, HDL1, and TMD8) significantly reduced cell growth, slowed proliferation, and induced marked apoptotic cell death.
CRISPR Negative Selection: In a cellular competitive transplant assay in immune-deficient mice, cell populations carrying CRISPR-Cas9-induced frame-shift mutations that knocked out DUSP2 expression rapidly died out, leaving only cells with in-frame 3-nucleotide deletions that preserved DUSP2 protein expression—proving that DUSP2 is strictly essential for lymphoma cell viability.
In Vivo Transgenic Validation: To confirm its oncogenic potential, they generated conditional knock-in mice crossed with CD19-Cre lines to selectively overexpress DUSP2 in B-cells. These transgenic mice displayed dramatic spleen enlargement (splenomegaly), rapid B-cell and T-cell hyper-proliferation, and accelerated malignant transformation.

Catalytic-Independent CDK1 Activation Machinery

Target Identification: Immunoaffinity purification followed by liquid chromatography-tandem mass spectrometry (IP-MS) identified CDK1 (cyclin-dependent kinase 1) as a primary DUSP2-associated protein in DLBCL cells.
The Activation Paradox: CDK1 activation requires the dephosphorylation of its inhibitory Tyr15 and Thr14 residues. DUSP2 overexpression reduced tyrosine-15 phosphorylation, whereas DUSP2 knockdown accumulated inhibitory phosphorylation, restricting cells in the G2/M transition.
Non-Catalytic Dephosphorylation: To test whether DUSP2 directly dephosphorylates CDK1, they generated catalytically dead mutants, including a C257S substitution and a catalytic-loop deletion. Surprisingly, these catalytically inactive mutants still successfully bound CDK1 and induced dephosphorylation of Tyr15.
CDC25 Scaffold Mechanism: DUSP2 contains an evolutionarily conserved structural motif (centered on residue D92) that physically recruits CDC25 phosphatases (specifically CDC25C). Rather than acting as the direct enzyme, DUSP2 serves as a physical scaffold that binds both CDK1 and CDC25C, facilitating the recruitment of CDC25C to catalyze CDK1 dephosphorylation, driving rapid mitotic cell cycle progression.

3. Assistant Professor Chatchai Phoomak

Department of Biology, Faculty of Science, Chulalongkorn University, Thailand

Presentation Topic: Uncovering N-Linked Glycosylation Drivers in Lung Cancer Through Whole-Genome CRISPR-Cas9 Screening

CRISPR-Cas9 Screen utilizing a Glycosylation Reporter

Pathology: Aberrant N-linked glycosylation is a major hallmark of cancer that regulates receptor tyrosine kinase signaling, protein folding, stability, and cellular homeostasis.
High-Throughput Screen: To systematically map the genetic regulators of glycosylation in lung cancer, the lab constructed a fluorescence-based HaloTag-1N glycosylation reporter. When the reporter protein is fully glycosylated, the fluorescent signal is blocked; under hypoglycosylation, a robust fluorescent signal is emitted.
Identified Drivers: Combining this reporter with a genome-wide CRISPR-Cas9 knockout screen in lung cancer cells, they identified key regulatory networks governing lipid-linked oligosaccharide biosynthesis, glucose metabolism, and endoplasmic reticulum (ER) quality control. Identified hits included GLUT1, UAP1, UGP2, ALG1, and components of the translocon-associated protein (TRAP) complex.

The GLUT1–UAP1–UGP2 Glycosylation Flux Axis

Glucose-Driven Aggressiveness: Cancer cells adapt their metabolism to high-glucose conditions. Spheroid and monolayer cultures under high glucose exhibited a marked upregulation of the GLUT1–UAP1–UGP2 metabolic axis, which increased intracellular UDP-GlcNAc pools and N-glycosylation flux.
Oncogenic Glycosylation Targets: This heightened glycosylation flux specifically targeted the cell-surface signaling receptor GP130, stabilizing it and activating downstream survival signaling pathways.
Therapeutic Vulnerability: Knocking out or pharmacologically inhibiting this axis disrupted GP130 N-linked glycosylation, triggered the unfolded protein response (UPR) / ER stress, induced cell-cycle arrest in the G1 phase, and significantly suppressed lung adenocarcinoma proliferation, migration, and invasiveness.
ALG1 Biomarker Core: In parallel, ALG1 (asparagine-linked glycosylation 1) was identified as a clinically overexpressed enzyme in lung cancer patients, with high ALG1 expression directly correlating with poor survival rates.

4. Dr. Nattawadee Panyain

Department of Biochemistry, Faculty of Medicine, Mahidol University, Thailand

Presentation Topic: From Probe to Target: Chemical Proteomics Approaches for Drug Target Discovery

Chemical Proteomics for Natural Product Target Deconvolution

The Unbiased Target Search: Natural products represent rich scaffolds for drug discovery, but identifying their direct physical targets inside living cells remains a bottleneck. The speaker's group developed an integrated chemical proteomics platform utilizing Activity-Based Protein Profiling (ABPP) and photo-crosslinking probes to map target networks under physiological conditions.

Curcumin Case Study in Viral and Oncogenic Contexts

Calu-3 Labeling: Using Calu-3 lung cancer cells, they synthesized a series of active curcumin probes. Concentration-dependent labeling assays revealed that curcumin does not target a single protein but binds to a multi-protein network spanning molecular weights from 25 kDa to 100 kDa.
Overlapping Targets: Quantitative mass spectrometry-based proteomics identified over 700 probe-binding proteins, with 546 targets consistently overlapping across different concentrations and viral infection states.
Mechanism of Viral Disruption: Pathway enrichment identified Reticulon-3 (RTN3) and Reticulon-4 (RTN4)—essential endoplasmic reticulum (ER)-shaping proteins—as direct physical binders of curcumin. During viral infection (such as SARS-CoV-2), viruses hijack the ER to form double-membrane vesicles (DMVs) to shelter their replication complexes. By binding RTN3 and RTN4, curcumin directly disrupts ER-membrane remodeling, halting DMV formation and blocking viral replication.

Cardamonin and Diazirine Photo-Probes

Cardamonin (NK49): Cardamonin is a natural chalcone isolated from fingerroot (Boesenbergia rotunda) with moderate anti-cancer properties. Through structural optimization, they synthesized a derivative, NK49, which exhibited significantly enhanced cytotoxic potency against A549 lung cancer cells. ABPP and pull-down proteomics mapped 18 shared targets and 65 unique targets of NK49 involved in critical cell-cycle and mitotic regulation.
Diazirine Reversible Targeting: For non-covalent, reversible natural products, they designed photo-reactive probes incorporating a diazirine group. Upon exposure to UV light, the diazirine group generates a highly reactive carbene that forms a permanent covalent bond with the bound target protein, allowing high-resolution mass spectrometry deconvolution of previously uncapturable drug-target complexes.

5. Ms. Sukasem Kiatsukasem

School of Biomolecular Science and Engineering, Vidyasirimedhi Institute of Science and Technology (VISTEC), Thailand

Presentation Topic: Tuning the efficiency of DNA amplification by dual-scaffolded catalytic condensates

Enhancing Isothermal Recombinase Polymerase Amplification (RPA)

Diagnostic Framework: Recombinase Polymerase Amplification (RPA) is a highly sensitive isothermal DNA amplification technique that has been integrated with CRISPR systems for rapid pathogen detection (such as COVID-19).
The Biomolecular Condensate Model: The RPA reaction relies on T4 bacteriophage-derived proteins, primarily the recombinase UvsX (transcribed phonetically as "USA") and the single-stranded binding protein gp32 (transcribed as "GP32"). These proteins form crowded, dual-scaffolded catalytic condensates that accelerate DNA-polymerase searching efficiency.
Truncation Dynamics: To understand the structural bottlenecks of this reaction, the researchers generated systematically truncated mutants of the recombinase UvsX (spanning mutants \(\Delta 6\) to \(\Delta 38\)).
Results: While extreme truncations (\(\Delta 38\) onward) abolished all catalytic and amplification activity, shorter truncations (specifically \(\Delta 6\) to \(\Delta 27\)) actually outperformed the wild-type recombinase, yielding significantly higher concentrations of target DNA amplicons. This proved that full-length recombinase enzyme activity is not the absolute bottleneck; rather, optimizing the biophysical condensate properties via targeted sequence truncation enhances diagnostic amplification efficiency.

6. Dr. Serap Pektaş

Department of Chemistry, Faculty of Arts and Sciences, Recep Tayyip Erdogan University, Turkey

Presentation Topic: Alternative Peptide Scaffolds for Modulation of the p53–MDM2 Axis: A Combined Computational and Cellular Study

Targeting the p53–MDM2 Interface with Non-Canonical Peptides

Clinical Need: In cancers that retain wild-type p53, the overexpression of its negative regulator MDM2 acts as a primary silencing mechanism, marking p53 for proteasomal degradation.
Structural Strategy: Traditional peptide-based inhibitors mimic the canonical alpha-helix of the p53 transactivation domain, utilizing a conserved hydrophobic triad. However, this study investigated whether non-canonical peptide scaffolds could successfully engage the MDM2 binding cleft.
Phosphorylation-Derived Generation: Alternative peptide candidates were computationally generated from phosphorylation-associated protein regions rather than p53 itself to evaluate how sequence context affects binding compatibility.
Intracellular HEK293T Expression: These designed candidate peptides were expressed intracellularly as stable fusion proteins in HEK293T cells. Quantitative western blot analysis demonstrated that several of these non-canonical peptide variants successfully stabilized and increased endogenous wild-type p53 protein levels.
Computational and MD Validation: Molecular dynamics (MD) simulations and structural docking confirmed that the designed active peptides achieved highly stable, alternative binding patterns inside the MDM2 cleft. They maintained the essential hydrophobic contacts while introducing novel flanking residues that improved compatibility with the MDM2 binding interface, establishing the feasibility of using non-canonical peptide scaffolds for p53 pathway reactivation.

1. Prof. Richard Simpson (Keynote Speaker)

Affiliation: La Trobe University, Melbourne, Australia
Presentation Title: Clinical relevance of extracellular vesicles in cancer — therapeutic and diagnostic potential
Key Content & Insights:
- Historical Perspective: Commenced by reflecting on the abstract book from the 2006 Thai Protein Society meeting, noting the extraordinary progress in protein chemistry, mass spectrometry, and young scientific talent in Thailand over the past 20 years.
- Extracellular Vesicles (EVs) & Intercellular Communication: Detailed how tumor cells release extracellular vesicles (exosomes) to interact with surrounding microenvironmental cells, driving disease progression.
- Proteomic Biomarker Discovery: Focused on mass spectrometry-based clinical proteomics to characterize EV cargo, facilitating the discovery of non-invasive diagnostic biomarkers and monitoring therapeutic response in oncology.

2. Prof. Chun-Jung Chen (Plenary Speaker)

Affiliation: National Synchrotron Radiation Research Center (NSRRC), Hsinchu, Taiwan
Presentation Title: Structural Biology of Viruses and Capsid Assembly
Key Content & Insights:
- Synchrotron & Cryo-EM Integration: Utilized single-particle Cryo-EM and synchrotron radiation facilities to resolve high-resolution viral structures and structural variations.
- Dynamics of Viral Capsid Assembly: Mapped viral particle assembly pathways, highlighting dynamic structural intermediates (such as T3 vs. T4-like sub-particles formed during assembly in the absence of viral genomic RNA).
- Infectious Disease Applications: Leveraged capsid architecture and native substrate-binding mechanics to design novel antiviral interventions and target viral infection machinery.

3. Assoc. Prof. Varodom Charoensawan (Invited Speaker)

Affiliation: Department of Biochemistry, Faculty of Science, Mahidol University, Thailand
Presentation Title: Characterisation of context-specific gene expression using multimodal transcriptomics
Key Content & Insights:
- Multimodal Omics Framework: Integrated bulk, single-cell, and spatially resolved transcriptomics with AI and computational modeling to map context-dependent gene regulatory networks.
- Human Cell Atlas Contributions: Discussed contributions to global initiatives like the Human Cell Atlas, building single-cell reference maps to define cellular diversity and disease states.
- Transcriptome vs. Proteome Uncoupling: Highlighted that discordance between transcript and protein levels reveals essential insights into cell plasticity, post-transcriptional regulation, and disease mechanisms.

4. Prof. Seiji Okada (Invited Speaker)

Affiliation: Joint Research Center for Human Retrovirus Infection, Kumamoto University, Japan
Presentation Title: Construction of a Novel Format of Bispecific T-cell Engager Antibodies Utilizing Cyclic Single-chain Antibodies and Analysis of Their Antitumor Effects
Key Content & Insights:
- Cyclic BiTE Design: Developed an engineered Bispecific T-cell Engager (BiTE) antibody platform utilizing cyclic single-chain antibodies to enhance structural stability and targeted cytotoxic T-cell engagement.
- Potent Anti-Tumor Efficacy: Demonstrated that cyclic BiTE constructs significantly reduce tumor burden, suppress metastasis, and inhibit tumor growth in primary cell models and patient-derived xenografts (PDX).
- Microenvironmental Versatility: Showed robust antitumor responses across varied tumor microenvironments, positioning the platform as a candidate for clinical cancer immunotherapies.

Source PST2026_005 contains the recording of the Day 2 Late-Morning Session (Protein Science for Precision Medicine and Sustainable Health) from the 20th International Symposium of the Protein Society of Thailand.

Here is the detailed extraction of the presentations featured in this video:

1. Prof. Waranyoo Phoolcharoen

Affiliation: Faculty of Pharmaceutical Sciences, Chulalongkorn University, Thailand
Presentation Topic: Green Biomanufacturing: Plant-Produced Antibodies as Next-Generation Biologics
Key Highlights & Methodology:
- Alternative Production Platform: Addressed the supply and cost bottlenecks of conventional monoclonal antibody manufacturing (which relies heavily on mammalian cell culture) by establishing plants as single-cell micro-reactors.
- Transient Expression System: Uses transient expression in Nicotiana benthamiana plants, offering rapid, flexible, and scalable production of therapeutic biologics.
- Economics in LMICs: Highlighted that while global manufacturing costs between plant and mammalian cells are debated, plant-based platforms offer dramatic cost reductions in low- and middle-income countries (LMICs), improving local healthcare accessibility.
- Expressing Complex Targets: Demonstrated that plant expression systems can successfully produce complex antibody formats that fail or express poorly in standard mammalian cell lines (e.g., CHO cells).

2. Prof. Yu-Ju Chen

Affiliation: Institute of Chemistry, Academia Sinica, Taiwan
Presentation Topic: Clinical Mass Spectrometry for Precision Oncology: From Proteome Discovery to Actionable Biomarkers
Key Highlights & Methodology:
- Translational Proteomics: Focuses on bridging large-scale proteomic profiling with clinical precision oncology.
- Mass Spectrometry Workflows: Advances high-throughput, quantitative mass spectrometry methodologies to analyze clinical cancer tissues and identify actionable biomarkers.
- Precision Therapeutics: Translates proteomic discoveries into personalized treatment strategies and higher-accuracy diagnostic tools for cancer patient care.

3. Dr. Tian (Sponsor Presentation by P. Intertrade Equipment Co., Ltd. / Thermo Fisher Scientific)

Title/Topic: Prep Smart, Discover More: Practical Sample Preparation Workflows for Proteomic Mass Spectrometry
Key Highlights & Methodology:
- Bottom-Up Proteomic Workflows: Focused on streamlining the multi-step bottom-up proteomic sample preparation pipeline—spanning protein extraction, denaturation, reduction, alkylation, enzymatic digestion, and peptide cleanup/desalting.
- Reducing Experimental Variability: Highlighted recent technological advancements and standardized sample prep kits designed to minimize human error and sample loss prior to mass spectrometry acquisition.

Extraction of PST2026_006

Session: Day 2 Afternoon Parallel Session — Mandarin B: Computational and AI-Driven Antibody Targeting and Epitope Mapping for Therapeutic Applications

1. Assoc. Prof. Duangdao Wichadakul

Affiliation: Department of Computer Engineering, Faculty of Engineering, Chulalongkorn University
Presentation Title: Towards an AI-Driven Platform for Peptide and Protein Engineering: From Generative Design to Property Prediction
Key Content & Highlights:
- Generative Sequence Design: Discusses deep learning and machine learning architectures for generating novelty-targeted peptide sequences and engineering functional proteins.
- Multi-Property Prediction Pipeline: Integrates predictive models for binding affinity, structural stability, and gas chromatography-mass spectrometry (GC/MS) retention times.
- Infrastructure & Collaboration: Outlines the lab’s data management systems, software tools, and collaborative frameworks for applying AI models to micro-peptides and therapeutic targets.

2. Prof. Chatchai Tayapiwatana

Affiliation: Faculty of Associated Medical Sciences, Chiang Mai University
Presentation Title: CD147 as a Target for Antibody Therapy Beyond Molecular Dynamics Simulations and AI Applications
Key Content & Highlights:
- Targeting Basigin (CD147): Focuses on CD147 as an essential receptor in cancer cell invasion, tissue remodeling, and viral entry.
- Biophysical Modeling: Combines molecular dynamics (MD) simulations with computational docking and machine learning to refine antibody binding interfaces and epitope specificity.
- Therapeutic Applications: Demonstrates how engineered antibody variants and nanobodies blocking CD147 suppress malignant progression.

3. Ms. Rissadee Phetpradit

Affiliation: Department of Biochemistry, Faculty of Science, Kasetsart University
Presentation Title: Biochemical and Structural Characterization of GH3 β-Glucosidase AtBgl3A from Acetivibrio Thermocellus
Key Content & Highlights:
- Enzyme Characterization: Explores the structural and catalytic properties of the glycoside hydrolase family 3 (GH3) β-glucosidase (AtBgl3A) isolated from the thermophilic bacterium Acetivibrio thermocellus.
- Lignocellulosic Biomass Processing: Highlights substrate specificity, kinetic parameters, and thermostability, evaluating its utility for industrial biomass bioconversion.

4. Asst. Prof. Patanachai Limpikirati

Affiliation: Faculty of Pharmaceutical Sciences, Chulalongkorn University
Presentation Title: DEPC Covalent Labeling Mass Spectrometry for Epitope Mapping of Therapeutic Antibody Targets
Key Content & Highlights:
- DEPC Covalent Labeling MS: Leverages diethyl pyrocarbonate (DEPC) modification coupled with mass spectrometry to map solvent accessibility and structural changes upon antibody binding.
- Residue-Level Resolution: Mapped conformational binding epitopes on therapeutic targets including PD-1 (identifying critical residues S27, S60, S62, S127, and K131 for nivolumab/CU-MAB) and the SARS-CoV-2 spike protein.

5. Dr. Kittikhun Wangkanont

Affiliation: Center of Excellence for Molecular Biology and Genomics of Shrimp, Department of Biochemistry, Faculty of Science, Chulalongkorn University
Presentation Title: Cryogenic Electron Microscopy Structure of the Chromoprotein Ovorubin
Key Content & Highlights:
- Cryo-EM Resolution: Applies single-particle cryogenic electron microscopy (Cryo-EM) to determine the 3D structure of ovorubin, a high-molecular-weight chromoprotein derived from snail eggs.
- Structural Function: Elucidates the oligomeric architecture, carotenoid/protein binding pocket assembly, and biophysical stability mechanisms.

6. Assoc. Prof. Kiattawee Choowongkomon

Affiliation: Department of Biochemistry, Faculty of Science, Kasetsart University
Presentation Title: Target, Block, Detect: A Multifaceted Approach to Antibody Application in Medicine and Agriculture
Key Content & Highlights:
- Engineered Antibody Formats: Reviews the development of single-chain variable fragments (scFvs), nanobodies (VHHs), and transbodies generated via phage display.
- Broad Applications: Demonstrates antibody applications ranging from receptor blocking in oncology (EGFR) and clinical diagnostics (NT-proBNP, NGAL) to veterinary diagnostics and agricultural pathogen detection.

7. Mr. Pirm Dhararag

Affiliation: Brandeis University & Chulalongkorn University
Presentation Title: Generative AI-Driven Design of Dual MDM2/MDMX Inhibitory Peptides with Integrated Drug-Likeness Optimization
Key Content & Highlights:
- Generative Peptide Design: Implements a computational pipeline utilizing generative AI to design dual-inhibitory peptides targeting both MDM2 and MDMX to restore wild-type p53 tumor suppressor activity.
- Multification & Modeling: Optimizes candidates using ADMET profiling, AlphaFold 3 complex prediction, and molecular dynamics simulations to ensure drug-likeness and binding stability.

1. Prof. Lindsay K. Pino (Keynote Speaker)

Affiliation: Talus Bioscience, Inc., Seattle, WA, USA
Presentation Title: Proteomics to Drive Drug Discovery Through Direct Measurement of Transcriptional Regulation
Key Content & Highlights:
- Drugging "Undruggable" Targets: Addresses nuclear transcription factors and chromatin regulators that lack traditional enzymatic binding pockets.
- Multiplexed Proteomic Assay: Uses a specialized physical separation and multiplexed proteomics workflow to directly measure real-time disruption of protein-DNA interactions in intact cells.
- Whole-Cell Off-Target Profiling: Evaluates whole-cell proteomic responses during hit triage to measure selectivity and off-target effects across the entire proteome.

2. Dr. Deanpen Japrung (Invited Speaker)

Affiliation: Director of Strategic Research Initiatives, National Nanotechnology Center (NANOTEC), NSTDA, Thailand
Presentation Title: The Single-Molecule Revolution: Decoding Protein Complexity and Sequencing with Nanopore Technology
Key Content & Highlights:
- Single-Molecule Nanopores: Reviews advancements in solid-state and biological nanopore sensing to analyze individual protein molecules and post-translational modifications (PTMs).
- Glycated Albumin Detection: Demonstrates nanopore-based discrimination of glycated albumin, where adding the glycan group increases current blockage to ~13%, decoded via fast-acting machine learning classifiers.
- Single-Molecule Conference: Announced the upcoming international conference on single-molecule protein sequencing hosted in Phuket, Thailand.

3. Prof. Siwanon Jirawatnotai (Invited Speaker)

Affiliation: Department of Pharmacology, Siriraj Hospital, Mahidol University
Presentation Title: Targeting KRAS Oncogenic Addiction in Cholangiocarcinoma
Key Content & Highlights:
- Cholangiocarcinoma Resistance: Focuses on overcoming severe chemoresistance in biliary tract cancers using patient-derived xenograft (PDX) and organoid models.
- KRAS Expression Dependency: Demonstrates that therapeutic responsiveness to KRAS pathway inhibitors correlates strongly with baseline KRAS expression levels rather than mutation status alone.
- Combination Therapies: Evaluates combination regimens pairing KRAS targeted agents with standard chemotherapy to block primary resistance.

4. Mr. Misal Rana Magar (Oral Presentation)

Affiliation: Institute of Molecular Biosciences, Mahidol University
Presentation Title: Enhancing Translation Efficiency via Combinatorial Design of Orthoflaviviral 5´ and 3´ Untranslated Regions
Key Content & Highlights:
- Chimeric UTR Library: Constructed a combinatorial library of 25 mRNA constructs pairing 5´ and 3´ untranslated regions (UTRs) from orthoflaviviruses (including Dengue, Zika, Yellow Fever, West Nile, and Japanese Encephalitis) driving NanoLuciferase.
- In Vitro & Cellular Testing: Transfected in vitro transcribed, capped mRNAs into human HEK293T cells, assessing luciferase expression at 24h and 72h.
- Top Candidates: Identified specific combinations—such as D5Y3 and Y5Y3—that yielded high translation efficiency and transcript stability, comparable to the human beta-globin (HBB) positive control.

5. Mr. Thada Wangsahnga (Oral Presentation)

Affiliation: Department of Biology, Faculty of Science, Mahidol University
Presentation Title: Production and Characterization of a Phage-Encoded LysM Protein with Preliminary Antibiofilm Activity Against Clostridioides difficile
Key Content & Highlights:
- Enzyme Therapy: Investigates phage-derived peptidoglycan hydrolases to combat hypervirulent, antibiotic-resistant Clostridioides difficile.
- Biophysical Characterization: Expressed and purified a phage LysM domain protein. Circular dichroism (CD) spectroscopy showed a predominantly beta-sheet secondary structure with thermal unfolding occurring at ~55°C.
- Antibiofilm Activity: Demonstrated significant preliminary antibiofilm and lytic activity against clinical C. difficile strains (strain 630 and R20291).