Periodic Reporting for period 1 - DRAIGON (Diagnosing Infections with Multi-Drug Resistant Microorganisms using AI-powered Genomic Antibiotic Susceptibility Prediction from Long-Read Sequencing Data)
Berichtszeitraum: 2024-01-01 bis 2025-06-30
1-Optimization of blood culture for human blood and synovial fluid as exemplary specimen types (WP2): DRAIGON advanced a standardized, automated workflow for rapid pathogen detection and genomic AST in blood and synovial fluid using long-read sequencing. Key steps included optimizing DNA extraction, host depletion, and integrating a stable ONT premix into the DRAIGONPrep microfluidic platform. This seamless workflow addresses urgent diagnostic needs and supports scalable adoption in clinical and public health settings.
2-Application of cloud-based analysis software using AI models for antibiotic susceptibility prediction (genomic AST; gAST)AMR marker detection and genomic surveillance (WP3/4/5/6): DRAIGON enhanced AI-powered tools for genomic profiling of bloodstream and joint infections by expanding the DraGNOME model for MIC prediction using long-read sequencing. Key work included refining pipelines, curating AMR/virulence databases, and training on 100,000+ genomes and >1000 in-house sequenced MDR strains. A unified cloud-based reporting tool was developed to support rapid, accurate diagnostics and AMR prediction, with potential impact in both high- and low-resource settings.
3-Clinical utility studies addressing blood stream infections (BSI) and prosthetic joint infections (PJI) (WP4/5/7) will be performed in two tertiary care institutions in the EU (Isala and OSS) and two university hospitals, one in the EU (UMCG) and one in the US (JHU), respectively: We advanced clinical validation of genomic pathogen profiling and rapid gAST in bloodstream and joint infections via multi-site studies (Isala, OSS, UMCG, JHU). DNA extraction protocols were optimized, and workflows tested on clinical samples within 24 hours of positivity. Proof-of-concept trials and meta-analyses support assessing diagnostic accuracy and patient impact, informing health-economic models and implementation in diverse healthcare settings.
4-Capacity-building and proof-of-concept study in a medium-resource environment (WP6).UHSN will use DRAIGONPrep and DraGNOME to carefully collect and analyze patient samples. The focus will be on detecting both common infections and those caused by bacteria that are resistant to multiple antibiotics. This will help us test how well the DraGNOME system works, including its cloud-based reporting tools. The goal is to gather strong evidence to show how the system can help doctors choose better antibiotics, fit smoothly into hospital routines, and improve patient care. This will benefit both hospitals with lots of resources and those with fewer resources. As part of this effort, we will also improve the labs’ capabilities by providing training and new equipment.
5-Development of automated liquid handling and sequencing scales for near point of care (WP2): After assessing several technologies, Camtech chose digital microfluidics (electrowetting-on-dielectric, EWOD) for its automated liquid handling prototype, developing a benchtop system and palm-sized microfluidic chips from conductive glass and plastic to perform DNA extraction and library preparation, incorporating clinical partners feedback; full performance testing is still ongoing.
6-Regulatory plan for IVDR acceptance(WP8): To get the device approved for the EU market under IVDR, we first define its purpose and risk category, and gap analysis. A quality system following ISO 13485 is ongoing and prepare all the necessary documents and test results. We also establish strong procedures for DraGNOME to monitor the device’s safety after launch, protect patient data under GDPR, ensure cybersecurity, and adjust for any national requirements. Throughout, we continuously track regulatory changes and assign clear roles and deadlines to secure CE marking and keep the device available and safe for patients.
2-Application of cloud-based analysis software using AI models for antibiotic susceptibility prediction (genomic AST; gAST)AMR marker detection and genomic surveillance (WP3/4/5/6): DRAIGON enhanced AI-powered tools for genomic profiling of bloodstream and joint infections by expanding the DraGNOME model for MIC prediction using long-read sequencing. Key work included refining pipelines, curating AMR/virulence databases, and training on 100,000+ genomes and >1000 in-house sequenced MDR strains. A unified cloud-based reporting tool was developed to support rapid, accurate diagnostics and AMR prediction, with potential impact in both high- and low-resource settings.
3-Clinical utility studies addressing blood stream infections (BSI) and prosthetic joint infections (PJI) (WP4/5/7) will be performed in two tertiary care institutions in the EU (Isala and OSS) and two university hospitals, one in the EU (UMCG) and one in the US (JHU), respectively: We advanced clinical validation of genomic pathogen profiling and rapid gAST in bloodstream and joint infections via multi-site studies (Isala, OSS, UMCG, JHU). DNA extraction protocols were optimized, and workflows tested on clinical samples within 24 hours of positivity. Proof-of-concept trials and meta-analyses support assessing diagnostic accuracy and patient impact, informing health-economic models and implementation in diverse healthcare settings.
4-Capacity-building and proof-of-concept study in a medium-resource environment (WP6).UHSN will use DRAIGONPrep and DraGNOME to carefully collect and analyze patient samples. The focus will be on detecting both common infections and those caused by bacteria that are resistant to multiple antibiotics. This will help us test how well the DraGNOME system works, including its cloud-based reporting tools. The goal is to gather strong evidence to show how the system can help doctors choose better antibiotics, fit smoothly into hospital routines, and improve patient care. This will benefit both hospitals with lots of resources and those with fewer resources. As part of this effort, we will also improve the labs’ capabilities by providing training and new equipment.
5-Development of automated liquid handling and sequencing scales for near point of care (WP2): After assessing several technologies, Camtech chose digital microfluidics (electrowetting-on-dielectric, EWOD) for its automated liquid handling prototype, developing a benchtop system and palm-sized microfluidic chips from conductive glass and plastic to perform DNA extraction and library preparation, incorporating clinical partners feedback; full performance testing is still ongoing.
6-Regulatory plan for IVDR acceptance(WP8): To get the device approved for the EU market under IVDR, we first define its purpose and risk category, and gap analysis. A quality system following ISO 13485 is ongoing and prepare all the necessary documents and test results. We also establish strong procedures for DraGNOME to monitor the device’s safety after launch, protect patient data under GDPR, ensure cybersecurity, and adjust for any national requirements. Throughout, we continuously track regulatory changes and assign clear roles and deadlines to secure CE marking and keep the device available and safe for patients.
WP1:The strategic, executive, and advisory bodies have been established and formalised, and the Coordinator is ensuring progress through effective monitoring, coordination, and reporting.
WP2 developed a standardized workflow for pathogen DNA extraction and modified library preparation for long-read sequencing from blood and synovial fluids. DNA extraction kits and library prep protocols were compared, and pre-enrichment conditions were optimized. Integration into the DRAIGONPrep microfluidic system was initiated. A proof-of-concept study and automation testing have started; protocols are in place and sequencing is ongoing, with troubleshooting underway. With the goal of developing a customized automation platform to bring next generation sequencing-based diagnostics closer to clinical use, Camtech’s work has focused on creating a suitable platform for this purpose. A digital microfluidic prototype, using electrowetting technology and controlled magnets, has been created with application specific chips. The chips are designed to perform the protocols, which are currently done manually. Once the performance of the prototype device is verified it will enable scientists to perform the analysis of patient derived bacterial samples in a faster and more automated fashion.
WP3 advanced tools for genomic pathogen profiling and gAST. Following initial delays, workflows for MIC prediction and ARG detection were developed, and the DraGNOME platform was initiated. Bioinformatic pipelines, resistance databases, and validation datasets were curated. Pathogen typing and outbreak tracing studies began. A first software version is available.
WP4 initiated evaluation of long-read sequencing workflows for bloodstream infections. Study protocols were approved, and comparative analyses of DNA extraction, concentration, and sequencing methods were conducted. A proof-of-concept study using optimized protocols showed promising results for pathogen ID and AMR gene detection. Further optimization is ongoing. Data collection and analysis workflows have been established.
WP5 initiated comparative studies to improve PJI diagnostics using long-read sequencing. Optimization of DNA extraction and pathogen enrichment workflows is ongoing at OSS and Isala, supported by pilot studies. Hybrid ONT/Illumina sequencing showed promising results. Protocols are in place and contacts established for continued clinical sample processing and data collection.
WP6 initiated capacity building for WGS-based AMR diagnostics at UHSN. Key achievements include staff training at UMCG, on-site consultation, and preparation of lab infrastructure. A detailed equipment list was developed, and procurement is underway. Protocols and SOPs were adapted with support from UMCG. Although administrative delays occurred, foundational steps were completed.
WP7 initiated key components of the health technology assessment (HTA). A study protocol and data plan were developed, and ethics approval obtained. The HTA Advisory Board met to validate the organizational survey and discuss economic modelling. A living systematic review protocol was registered (PROSPERO), with screening completed.
WP8 initiated preparatory steps toward IVDR compliance for DRAIGONPrep and DraGNOME. A QMS framework (ISO 13485) was outlined, and initial regulatory discussions were held with Camtech and potential CRO partners. Plans are available for device classification, gap analysis, and GDPR compliance. Contacts are in place to support next steps, including aligning with evolving EU AI and medical device regulations.
WP2 developed a standardized workflow for pathogen DNA extraction and modified library preparation for long-read sequencing from blood and synovial fluids. DNA extraction kits and library prep protocols were compared, and pre-enrichment conditions were optimized. Integration into the DRAIGONPrep microfluidic system was initiated. A proof-of-concept study and automation testing have started; protocols are in place and sequencing is ongoing, with troubleshooting underway. With the goal of developing a customized automation platform to bring next generation sequencing-based diagnostics closer to clinical use, Camtech’s work has focused on creating a suitable platform for this purpose. A digital microfluidic prototype, using electrowetting technology and controlled magnets, has been created with application specific chips. The chips are designed to perform the protocols, which are currently done manually. Once the performance of the prototype device is verified it will enable scientists to perform the analysis of patient derived bacterial samples in a faster and more automated fashion.
WP3 advanced tools for genomic pathogen profiling and gAST. Following initial delays, workflows for MIC prediction and ARG detection were developed, and the DraGNOME platform was initiated. Bioinformatic pipelines, resistance databases, and validation datasets were curated. Pathogen typing and outbreak tracing studies began. A first software version is available.
WP4 initiated evaluation of long-read sequencing workflows for bloodstream infections. Study protocols were approved, and comparative analyses of DNA extraction, concentration, and sequencing methods were conducted. A proof-of-concept study using optimized protocols showed promising results for pathogen ID and AMR gene detection. Further optimization is ongoing. Data collection and analysis workflows have been established.
WP5 initiated comparative studies to improve PJI diagnostics using long-read sequencing. Optimization of DNA extraction and pathogen enrichment workflows is ongoing at OSS and Isala, supported by pilot studies. Hybrid ONT/Illumina sequencing showed promising results. Protocols are in place and contacts established for continued clinical sample processing and data collection.
WP6 initiated capacity building for WGS-based AMR diagnostics at UHSN. Key achievements include staff training at UMCG, on-site consultation, and preparation of lab infrastructure. A detailed equipment list was developed, and procurement is underway. Protocols and SOPs were adapted with support from UMCG. Although administrative delays occurred, foundational steps were completed.
WP7 initiated key components of the health technology assessment (HTA). A study protocol and data plan were developed, and ethics approval obtained. The HTA Advisory Board met to validate the organizational survey and discuss economic modelling. A living systematic review protocol was registered (PROSPERO), with screening completed.
WP8 initiated preparatory steps toward IVDR compliance for DRAIGONPrep and DraGNOME. A QMS framework (ISO 13485) was outlined, and initial regulatory discussions were held with Camtech and potential CRO partners. Plans are available for device classification, gap analysis, and GDPR compliance. Contacts are in place to support next steps, including aligning with evolving EU AI and medical device regulations.