Charles Y. Tan
Charles Y. Tan has a total of 102 co-authors. Here are the top co-authors by number of shared publications:
- Alexander Muik
BioNTech SE, Mainz, Germany
1 shared publication - Alptekin Gueler
Biontech RNA Pharmaceuticals GmbH
1 shared publication - Andreas N. Kuhn
Biontech RNA Pharmaceuticals GmbH
1 shared publication - André P. Heinen
Biontech RNA Pharmaceuticals GmbH
1 shared publication - Ann-Kathrin Wallisch
Biontech RNA Pharmaceuticals GmbH
1 shared publication - Annette B. Vogel
BioNTech, Mainz, Germany
1 shared publication
Latest Publications
- Open Access
- 29 October 2026
- Published by Wiley in Clinical and Translational Science
- Vol. 19 (11) , e70747
Compliance by Design: Governing AI ‐Augmented Quantitative Analysis in Clinical and Translational Research
Artificial intelligence (AI) is moving from promise to routine practice in clinical and translational research, and large language models (LLMs) are now used to scope analyses, draft and debug code, and accelerate the quantitative work that supports clinical and translational decisions. This adoption is outpacing the operational safeguards that keep such work reproducible and auditable, and the central risk is not only erroneous code but the quieter erosion of the analytical discipline on which trustworthy bench‐to‐bedside decisions depend. In this position paper, we argue that AI can be integrated into regulated quantitative workflows without sacrificing reproducibility, provided it operates inside a compliance‐by‐design architecture built on two principles: analytical intent must be formally encoded and locked before execution, and the analysis itself must run outside the LLM as specification‐verified code in a validated environment. We describe three failure modes that ordinary quality control does not detect, namely context drift, the displacement effect, and implementation drift, and we link each to documented model behavior. For open‐ended investigation, these principles are operationalized as a four‐step execution firewall; for recurring development studies such as assay validations and animal safety studies, they are operationalized through a governed library of documented, validated agents, supported by fit‐for‐purpose validation, a clear separation of human and machine roles, and post‐deployment monitoring. Both approaches can be implemented with components available today. Used this way, AI enhances rather than replaces quantitative expertise, which now resides less in the code a model produces than in the constraints a scientist defines before it runs.
- 29 October 2026
- Published by Wiley in Clinical and Translational Science
- Vol. 19 (11) , e70747
Open AccessCompliance by Design: Governing AI ‐Augmented Quantitative Analysis in Clinical and Translational Research
Artificial intelligence (AI) is moving from promise to routine practice in clinical and translational research, and large language models (LLMs) are now used to scope analyses, draft and debug code, and accelerate the quantitative work that supports clinical and translational decisions. This adoption is outpacing the operational safeguards that keep such work reproducible and auditable, and the central risk is not only erroneous code but the quieter erosion of the analytical discipline on which trustworthy bench‐to‐bedside decisions depend. In this position paper, we argue that AI can be integrated into regulated quantitative workflows without sacrificing reproducibility, provided it operates inside a compliance‐by‐design architecture built on two principles: analytical intent must be formally encoded and locked before execution, and the analysis itself must run outside the LLM as specification‐verified code in a validated environment. We describe three failure modes that ordinary quality control does not detect, namely context drift, the displacement effect, and implementation drift, and we link each to documented model behavior. For open‐ended investigation, these principles are operationalized as a four‐step execution firewall; for recurring development studies such as assay validations and animal safety studies, they are operationalized through a governed library of documented, validated agents, supported by fit‐for‐purpose validation, a clear separation of human and machine roles, and post‐deployment monitoring. Both approaches can be implemented with components available today. Used this way, AI enhances rather than replaces quantitative expertise, which now resides less in the code a model produces than in the constraints a scientist defines before it runs.
- 20 November 2025
- Published by Wiley in Statistics in Medicine
- Vol. 44 (25-27) , e70315
An Information Criterion Approach for Assessing Agreement When Comparing Two Methods of Measurement
In laboratory and clinical sciences, it is a common enough challenge to compare two methods of measurement on the same set of samples or closely related samples with the goal of assessing agreement. Current regulatory guidance cites the Bland–Altman plot, Deming regression, and concordance correlation coefficient. Each choice relies on its assumption of a particular model. A new statistical approach based on an information criterion is proposed. This integrated approach evaluates the data against six models simultaneously. This information criterion approach provides an objective, data‐driven, easily calculated, informative, and clear decision rule. Real data sets from the assay comparison of patient‐centric sampling are used to illustrate the new approach.
- 20 November 2025
- Published by Wiley in Statistics in Medicine
- Vol. 44 (25-27) , e70315
An Information Criterion Approach for Assessing Agreement When Comparing Two Methods of Measurement
In laboratory and clinical sciences, it is a common enough challenge to compare two methods of measurement on the same set of samples or closely related samples with the goal of assessing agreement. Current regulatory guidance cites the Bland–Altman plot, Deming regression, and concordance correlation coefficient. Each choice relies on its assumption of a particular model. A new statistical approach based on an information criterion is proposed. This integrated approach evaluates the data against six models simultaneously. This information criterion approach provides an objective, data‐driven, easily calculated, informative, and clear decision rule. Real data sets from the assay comparison of patient‐centric sampling are used to illustrate the new approach.
- Open Access
- 30 May 2025
- Published by Elsevier in Diagnostic Microbiology and Infectious Disease
- Vol. 113 (2) , 116920
Interlaboratory study to assess precision and reproducibility of the meningococcal antigen surface expression (MEASURE) assay to quantify factor H binding protein expression at the surface of meningococcal serogroup B strains
- 30 May 2025
- Published by Elsevier in Diagnostic Microbiology and Infectious Disease
- Vol. 113 (2) , 116920
Open AccessInterlaboratory study to assess precision and reproducibility of the meningococcal antigen surface expression (MEASURE) assay to quantify factor H binding protein expression at the surface of meningococcal serogroup B strains
- Open Access
A prefusion SARS-CoV-2 spike RNA vaccine is highly immunogenic and prevents lung infection in non-human primates
To contain the coronavirus disease 2019 (COVID-19) pandemic, a safe and effective vaccine against the new severe acute respiratory syndrome coronavirus-2 (SARS-CoV-2) is urgently needed in quantities sufficient to immunise large populations. In this study, we report the design, preclinical development, immunogenicity and anti-viral protective effect in rhesus macaques of the BNT162b2 vaccine candidate. BNT162b2 contains an LNP-formulated nucleoside-modified mRNA that encodes the spike glycoprotein captured in its prefusion conformation. After expression of the BNT162b2 coding sequence in cells, approximately 20% of the spike molecules are in the one-RBD ‘up’, two-RBD ‘down’ state. Immunisation of mice with a single dose of BNT162b2 induced dose level-dependent increases in pseudovirus neutralisation titers. Prime-boost vaccination of rhesus macaques elicited authentic SARS-CoV-2 neutralising geometric mean titers 10.2 to 18.0 times that of a SARS-CoV-2 convalescent human serum panel. BNT162b2 generated strong T H 1 type CD4 + and IFNγ + CD8 + T-cell responses in mice and rhesus macaques. The BNT162b2 vaccine candidate fully protected the lungs of immunised rhesus macaques from infectious SARS-CoV-2 challenge. BNT162b2 is currently being evaluated in a global, pivotal Phase 2/3 trial ( NCT04368728 ).
Open AccessA prefusion SARS-CoV-2 spike RNA vaccine is highly immunogenic and prevents lung infection in non-human primates
To contain the coronavirus disease 2019 (COVID-19) pandemic, a safe and effective vaccine against the new severe acute respiratory syndrome coronavirus-2 (SARS-CoV-2) is urgently needed in quantities sufficient to immunise large populations. In this study, we report the design, preclinical development, immunogenicity and anti-viral protective effect in rhesus macaques of the BNT162b2 vaccine candidate. BNT162b2 contains an LNP-formulated nucleoside-modified mRNA that encodes the spike glycoprotein captured in its prefusion conformation. After expression of the BNT162b2 coding sequence in cells, approximately 20% of the spike molecules are in the one-RBD ‘up’, two-RBD ‘down’ state. Immunisation of mice with a single dose of BNT162b2 induced dose level-dependent increases in pseudovirus neutralisation titers. Prime-boost vaccination of rhesus macaques elicited authentic SARS-CoV-2 neutralising geometric mean titers 10.2 to 18.0 times that of a SARS-CoV-2 convalescent human serum panel. BNT162b2 generated strong T H 1 type CD4 + and IFNγ + CD8 + T-cell responses in mice and rhesus macaques. The BNT162b2 vaccine candidate fully protected the lungs of immunised rhesus macaques from infectious SARS-CoV-2 challenge. BNT162b2 is currently being evaluated in a global, pivotal Phase 2/3 trial ( NCT04368728 ).
- 3 January 2020
- Published by Springer Nature in The AAPS Journal
- Vol. 22 (2) , 19
Criteria to Reevaluate Anti-drug Antibody Assay Cut Point Suitability in the Target Population
After tier 1 and 2 cut points for anti-drug antibody (ADA) assays are derived during pre-study assay validation in a population, there is a need to verify the continued appropriateness of the...
- 3 January 2020
- Published by Springer Nature in The AAPS Journal
- Vol. 22 (2) , 19
Criteria to Reevaluate Anti-drug Antibody Assay Cut Point Suitability in the Target Population
After tier 1 and 2 cut points for anti-drug antibody (ADA) assays are derived during pre-study assay validation in a population, there is a need to verify the continued appropriateness of the...
- Open Access
- 29 August 2018
- Published by American Society for Microbiology in mSphere
- Vol. 3 (4) , e00127-18
Evaluation of a Validated Luminex-Based Multiplex Immunoassay for Measuring Immunoglobulin G Antibodies in Serum to Pneumococcal Capsular Polysaccharides
The pneumococcal enzyme-linked immunosorbent assay (ELISA) measures IgG antibodies in human serum, and it is an important assay that supports licensure of pneumococcal vaccines. The immune correlate of protection, 0.35 µg/ml of IgG antibodies, was determined by the ELISA method. Pfizer has developed a new Luminex-based assay platform to replace the ELISA. These papers describe the important work of (i) validating the Luminex-based assay and (ii) bridging the immune correlate of protection (0.35 µg/ml IgG) to equivalent values reported by the Luminex platform.
- 29 August 2018
- Published by American Society for Microbiology in mSphere
- Vol. 3 (4) , e00127-18
Open AccessEvaluation of a Validated Luminex-Based Multiplex Immunoassay for Measuring Immunoglobulin G Antibodies in Serum to Pneumococcal Capsular Polysaccharides
The pneumococcal enzyme-linked immunosorbent assay (ELISA) measures IgG antibodies in human serum, and it is an important assay that supports licensure of pneumococcal vaccines. The immune correlate of protection, 0.35 µg/ml of IgG antibodies, was determined by the ELISA method. Pfizer has developed a new Luminex-based assay platform to replace the ELISA. These papers describe the important work of (i) validating the Luminex-based assay and (ii) bridging the immune correlate of protection (0.35 µg/ml IgG) to equivalent values reported by the Luminex platform.
- Open Access
- 1 November 2015
- Published by American Society for Microbiology in Clinical and Vaccine Immunology
- Vol. 22 (11) , 1154-1159
Assignment of Weight-Based Antibody Units for Seven Additional Serotypes to a Human Pneumococcal Standard Reference Serum, 007sp
The pneumococcal enzyme-linked immunosorbent assay (ELISA) reference standard serum, lot 89SF, has been in use since 1990 and was replaced in 2013 with a new reference standard, 007sp, that is projected to be available for the next 25 years. 007sp was generated under an FDA-approved clinical protocol; 278 adult volunteers were immunized with the 23-valent unconjugated polysaccharide vaccine Pneumovax II, and a unit of blood was obtained twice from each immunized subject within 120 days following immunization. Pooled serum was prepared from the plasma of 262 subjects, filled at 6 ml per vial, and lyophilized. Five independent laboratories participated in bridging the serotype-specific IgG assignments for 89SF to the new reference standard, 007sp, to establish equivalent reference values for 13 pneumococcal capsular serotypes (1,3, 4, 5, 6A, 6B, 7F, 9V, 14, 18C, 19A, 19F, and 23F) by using the WHO reference ELISA. In a second study involving three laboratories, a similar protocol was used to assign weight-based IgG concentrations in micrograms per ml to 007sp of seven serotypes (8, 10A, 11A, 12F, 15B, 22F, and 33F) also present in the 23-valent pneumococcal unconjugated polysaccharide vaccine. In addition, the IgG assignments for a 12-member WHO quality control (QC) serum panel were also extended to cover these seven serotypes. Agreement was excellent, with a concordance correlation coefficient ( r c ) of >0.996 when each laboratory was compared to the assigned values for the 12 WHO QC serum samples. There are four remaining pneumococcal serotypes (2, 9N, 17F, and 20) found in Pneumovax II for which IgG assignments exist for 89SF and remain to be bridged.
- 1 November 2015
- Published by American Society for Microbiology in Clinical and Vaccine Immunology
- Vol. 22 (11) , 1154-1159
Open AccessAssignment of Weight-Based Antibody Units for Seven Additional Serotypes to a Human Pneumococcal Standard Reference Serum, 007sp
The pneumococcal enzyme-linked immunosorbent assay (ELISA) reference standard serum, lot 89SF, has been in use since 1990 and was replaced in 2013 with a new reference standard, 007sp, that is projected to be available for the next 25 years. 007sp was generated under an FDA-approved clinical protocol; 278 adult volunteers were immunized with the 23-valent unconjugated polysaccharide vaccine Pneumovax II, and a unit of blood was obtained twice from each immunized subject within 120 days following immunization. Pooled serum was prepared from the plasma of 262 subjects, filled at 6 ml per vial, and lyophilized. Five independent laboratories participated in bridging the serotype-specific IgG assignments for 89SF to the new reference standard, 007sp, to establish equivalent reference values for 13 pneumococcal capsular serotypes (1,3, 4, 5, 6A, 6B, 7F, 9V, 14, 18C, 19A, 19F, and 23F) by using the WHO reference ELISA. In a second study involving three laboratories, a similar protocol was used to assign weight-based IgG concentrations in micrograms per ml to 007sp of seven serotypes (8, 10A, 11A, 12F, 15B, 22F, and 33F) also present in the 23-valent pneumococcal unconjugated polysaccharide vaccine. In addition, the IgG assignments for a 12-member WHO quality control (QC) serum panel were also extended to cover these seven serotypes. Agreement was excellent, with a concordance correlation coefficient ( r c ) of >0.996 when each laboratory was compared to the assigned values for the 12 WHO QC serum samples. There are four remaining pneumococcal serotypes (2, 9N, 17F, and 20) found in Pneumovax II for which IgG assignments exist for 89SF and remain to be bridged.
- 1 August 1999
- Published by JSTOR in Technometrics
- Vol. 41 (3) , 192
Measurement-Methods Comparisons and Linear Statistical Relationship
The linear statistical relationship model can be used to compare two measurement procedures. Crude methods are now typically used in practice because the statistical concepts are often not understood by practitioners. To clarify concepts, we define three types of equivalency based on accuracy and precision of the measurement methods and show that the confidence interval for slope depends on sample size and information size. These insights lead to obtaining bounded confidence intervals and equivalence tests. The derived results provide the basis for designing and analyzing a comparison experiment. An example from practice is used to illustrate the methodology.
- 1 August 1999
- Published by JSTOR in Technometrics
- Vol. 41 (3) , 192
Measurement-Methods Comparisons and Linear Statistical Relationship
The linear statistical relationship model can be used to compare two measurement procedures. Crude methods are now typically used in practice because the statistical concepts are often not understood by practitioners. To clarify concepts, we define three types of equivalency based on accuracy and precision of the measurement methods and show that the confidence interval for slope depends on sample size and information size. These insights lead to obtaining bounded confidence intervals and equivalence tests. The derived results provide the basis for designing and analyzing a comparison experiment. An example from practice is used to illustrate the methodology.