Kam Hon Hoi
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Kam Hon Hoi has a total of 194 co-authors. Here are the top co-authors by number of shared publications:
- George Georgiou
Department of Molecular Biosciences, The University of Texas at Austin, Austin, TX, USA
6 shared publications - Gregory C. Ippolito
Texas Biomedical Research Institute
5 shared publications - Edward M. Marcotte
Department of Molecular Biosciences, The University of Texas at Austin, Austin, TX 78712, USA
3 shared publications - Isidro Hötzel
Department of Antibody Engineering, Genentech, South San Francisco, California, USA
3 shared publications Jason J. Lavinder
Department of Chemical Engineering, Institute for Cellular and Molecular Biology, Section of Molecular Genetics and Microbiology, Center for Systems and Synthetic Biology, Department of Biomedical Engineering, Department of Chemistry and Biochemistry, and Applied Research Laboratories, University of Texas at Austin, Austin, TX 78712-1062
3 shared publicationsYariv Wine
Department of Chemical Engineering
3 shared publications
Latest Publications
- Open Access
Learning gene co-regulation across cells with Multicellular Transformer
Understanding how genes coordinate across cells is central to explaining tissue function and disease, yet most existing models still take a single cell as input and therefore emphasize cell-intrinsic regulation. We present the Multicellular Transformer (MCT), a foundation model built around a different assumption: cells function together within tissues, so the appropriate unit of input is not one cell in isolation but a set of cells from the same sample. By jointly modeling multiple cells, MCT learns context-aware gene embeddings that capture gene-gene interactions across the tissue environment. Trained on a large set of tissue samples (>4200) across more than 30 tissue types, MCT recovers ligand-receptor interactions, infers cross-cell signaling programs, and reveals spatial co-localization patterns directly from scRNA-seq data. Its contextual embeddings identify disease-specific rewiring of gene neighborhoods in systemic lupus erythematosus and COVID-19 and predict population-level perturbation effects in cancer without additional training. By linking gene-level regulation to tissue-scale organization, MCT provides a scalable framework for modeling multicellular biology and enables in silico discovery of therapeutic targets and tissue-specific dependencies.
Open AccessLearning gene co-regulation across cells with Multicellular Transformer
Understanding how genes coordinate across cells is central to explaining tissue function and disease, yet most existing models still take a single cell as input and therefore emphasize cell-intrinsic regulation. We present the Multicellular Transformer (MCT), a foundation model built around a different assumption: cells function together within tissues, so the appropriate unit of input is not one cell in isolation but a set of cells from the same sample. By jointly modeling multiple cells, MCT learns context-aware gene embeddings that capture gene-gene interactions across the tissue environment. Trained on a large set of tissue samples (>4200) across more than 30 tissue types, MCT recovers ligand-receptor interactions, infers cross-cell signaling programs, and reveals spatial co-localization patterns directly from scRNA-seq data. Its contextual embeddings identify disease-specific rewiring of gene neighborhoods in systemic lupus erythematosus and COVID-19 and predict population-level perturbation effects in cancer without additional training. By linking gene-level regulation to tissue-scale organization, MCT provides a scalable framework for modeling multicellular biology and enables in silico discovery of therapeutic targets and tissue-specific dependencies.
- Open Access
Inferring cellular communication through mapping cells in space using Tangram 2
Cell-to-cell communication (CCC) shapes development, immunity, and disease, yet current spatial transcriptomics (SRT) platforms rarely achieve both single-cell resolution and high-quality transcriptome-wide coverage to accurately characterize CCC in the tissue microenvironment. Tangram2 bridges this gap by integrating single-cell RNA sequencing (scRNA-seq) with SRT to identify genes whose expression changes as a function of neighboring cell types. By accurately mapping cells in tissue space, Tangram2 disentangles interaction-driven transcriptional shifts from intrinsic identity markers, yielding mechanistic CCC maps. Validation across diverse settings-including Slide-tags, co-cultured experiments, human lymph nodes, and simulations-demonstrates high accuracy. Applied to triple-negative breast cancer (TNBC) and cutaneous squamous cell carcinoma (cSCC), Tangram2 recapitulates known biology and uncovers new hypotheses, including various immunosuppressive mechanisms in TNBC and a macrophage-regulatory T-cell circuit associated with survival in cSCC.
Open AccessInferring cellular communication through mapping cells in space using Tangram 2
Cell-to-cell communication (CCC) shapes development, immunity, and disease, yet current spatial transcriptomics (SRT) platforms rarely achieve both single-cell resolution and high-quality transcriptome-wide coverage to accurately characterize CCC in the tissue microenvironment. Tangram2 bridges this gap by integrating single-cell RNA sequencing (scRNA-seq) with SRT to identify genes whose expression changes as a function of neighboring cell types. By accurately mapping cells in tissue space, Tangram2 disentangles interaction-driven transcriptional shifts from intrinsic identity markers, yielding mechanistic CCC maps. Validation across diverse settings-including Slide-tags, co-cultured experiments, human lymph nodes, and simulations-demonstrates high accuracy. Applied to triple-negative breast cancer (TNBC) and cutaneous squamous cell carcinoma (cSCC), Tangram2 recapitulates known biology and uncovers new hypotheses, including various immunosuppressive mechanisms in TNBC and a macrophage-regulatory T-cell circuit associated with survival in cSCC.
- 15 June 2022
- Published by Oxford University Press (OUP) in The Journal of Immunology
- Vol. 208 (12) , 2632-2642
Activation-Induced Cytidine Deaminase Impacts the Primary Antibody Repertoire in Naive Mice
Genetic and environmental cues shape the evolution of the B cell Ig repertoire. Activation-induced cytidine deaminase (AID) is essential to generating Ig diversity through isotype class switching and somatic mutations, which then directly influence clonal selection. Impaired B cell development in AID-knockout mice has made it difficult to study Ig diversification in an aging repertoire. Therefore, in this report, we used a novel inducible AID-knockout mouse model and discovered that deleting AID in adult mice caused spontaneous germinal center formation. Deep sequencing of the IgH repertoire revealed that Ab diversification begins early in life and evolves over time. Our data suggest that activated B cells form germinal centers at steady state and facilitate continuous diversification of the B cell repertoire. In support, we identified shared B cell lineages that were class switched and showed age-dependent rates of mutation. Our data provide novel context to the genesis of the B cell repertoire that may benefit the understanding of autoimmunity and the strength of an immune response to infection.
- 15 June 2022
- Published by Oxford University Press (OUP) in The Journal of Immunology
- Vol. 208 (12) , 2632-2642
Activation-Induced Cytidine Deaminase Impacts the Primary Antibody Repertoire in Naive Mice
Genetic and environmental cues shape the evolution of the B cell Ig repertoire. Activation-induced cytidine deaminase (AID) is essential to generating Ig diversity through isotype class switching and somatic mutations, which then directly influence clonal selection. Impaired B cell development in AID-knockout mice has made it difficult to study Ig diversification in an aging repertoire. Therefore, in this report, we used a novel inducible AID-knockout mouse model and discovered that deleting AID in adult mice caused spontaneous germinal center formation. Deep sequencing of the IgH repertoire revealed that Ab diversification begins early in life and evolves over time. Our data suggest that activated B cells form germinal centers at steady state and facilitate continuous diversification of the B cell repertoire. In support, we identified shared B cell lineages that were class switched and showed age-dependent rates of mutation. Our data provide novel context to the genesis of the B cell repertoire that may benefit the understanding of autoimmunity and the strength of an immune response to infection.
- Open Access
- 2 June 2022
- Published by National Academy of Sciences in Proceedings of the National Academy of Sciences
- Vol. 119 (23)
Antibody homotypic interactions are encoded by germline light chain complementarity determining region 2
Significance: Weak transient interactions are fundamental to immune responses, enabling avidity-driven triggers for pathogen neutralization and cellular regulation. In contrast to obligate binding interactions that can be directly investigated structurally, the low or transitory abundance of weak interactions make them difficult to identify and characterize. This study leverages receptor agonism systems that are sensitive to oligomerization to investigate transient homotypic interfaces between antibody Fab regions. Our results show that self-association determinants are encoded naturally by the antibody germline through light chain complementarity determining region 2 (CDRL2), and these determinants can be engineered into antibodies to enhance their therapeutic properties. Insights into avidity-driven interactions create opportunities for optimization, and accordingly this work expands the engineering toolbox for antibody-based drugs.
- 2 June 2022
- Published by National Academy of Sciences in Proceedings of the National Academy of Sciences
- Vol. 119 (23)
Open AccessAntibody homotypic interactions are encoded by germline light chain complementarity determining region 2
Significance: Weak transient interactions are fundamental to immune responses, enabling avidity-driven triggers for pathogen neutralization and cellular regulation. In contrast to obligate binding interactions that can be directly investigated structurally, the low or transitory abundance of weak interactions make them difficult to identify and characterize. This study leverages receptor agonism systems that are sensitive to oligomerization to investigate transient homotypic interfaces between antibody Fab regions. Our results show that self-association determinants are encoded naturally by the antibody germline through light chain complementarity determining region 2 (CDRL2), and these determinants can be engineered into antibodies to enhance their therapeutic properties. Insights into avidity-driven interactions create opportunities for optimization, and accordingly this work expands the engineering toolbox for antibody-based drugs.
- Open Access
- 9 August 2019
- Published by Springer Nature in Communications Biology
- Vol. 2 (1) , 1-10
Massively parallel single-cell B-cell receptor sequencing enables rapid discovery of diverse antigen-reactive antibodies
Obtaining full-length antibody heavy- and light-chain variable regions from individual B cells at scale remains a challenging problem. Here we use high-throughput single-cell B-cell receptor sequencing (scBCR-seq) to obtain accurately paired full-length variable regions in a massively parallel fashion. We sequenced more than 250,000 B cells from rat, mouse and human repertoires to characterize their lineages and expansion. In addition, we immunized rats with chicken ovalbumin and profiled antigen-reactive B cells from lymph nodes of immunized animals. The scBCR-seq data recovered 81% (n = 56/69) of B-cell lineages identified from hybridomas generated from the same set of B cells subjected to scBCR-seq. Importantly, scBCR-seq identified an additional 710 candidate lineages not recovered as hybridomas. We synthesized, expressed and tested 93 clones from the identified lineages and found that 99% (n = 92/93) of the clones were antigen-reactive. Our results establish scBCR-seq as a powerful tool for antibody discovery.
- 9 August 2019
- Published by Springer Nature in Communications Biology
- Vol. 2 (1) , 1-10
Open AccessMassively parallel single-cell B-cell receptor sequencing enables rapid discovery of diverse antigen-reactive antibodies
Obtaining full-length antibody heavy- and light-chain variable regions from individual B cells at scale remains a challenging problem. Here we use high-throughput single-cell B-cell receptor sequencing (scBCR-seq) to obtain accurately paired full-length variable regions in a massively parallel fashion. We sequenced more than 250,000 B cells from rat, mouse and human repertoires to characterize their lineages and expansion. In addition, we immunized rats with chicken ovalbumin and profiled antigen-reactive B cells from lymph nodes of immunized animals. The scBCR-seq data recovered 81% (n = 56/69) of B-cell lineages identified from hybridomas generated from the same set of B cells subjected to scBCR-seq. Importantly, scBCR-seq identified an additional 710 candidate lineages not recovered as hybridomas. We synthesized, expressed and tested 93 clones from the identified lineages and found that 99% (n = 92/93) of the clones were antigen-reactive. Our results establish scBCR-seq as a powerful tool for antibody discovery.
- Open Access
Heavy Chain CDR 3 and Junctional Length Biases in Human Antibody Repertoires Associated with Heavy and Light Chain Germline Utilization
Antibody variable domain sequence diversity is generated by recombination of germline segments. The third complementarity-determining region of the heavy chain (CDR H3) is the region of highest sequence diversity and is formed by the joining of heavy chain V H , D H and J H germline segments combined with random nucleotide trimming and additions between these segments. We show that CDR H3 length distribution is biased in human antibody repertoires as a function of V H , V L and J H germline segment utilization. Most length biases are apparent in the naïve B cell compartment, with a significant bias towards shorter CDR H3 sequences observed in association with a subset of V H and V L germlines in the antigen experienced compartment. Similar biases were not observed in nonproductive heavy chain recombination products, indicating selection of the repertoire during B cell maturation as a major driver of the length biases. Some V H -associated CDR H3 length biases are dependent on utilization of specific J H germline segments in a manner not directly linked to J H segment length in the germline, but are rather associated with selection of differentially trimmed J H segments in the naïve compartment. In addition, D H segment and N-region random nucleotide insertion lengths within CDR H3 in the naïve compartment were also biased by specific V H /J H germline combinations, indicating a complex set of constraints between germline segments selected during repertoire maturation. Our findings reveal biases in the antibody diversity landscape shaped by V H , V L , and J H germline features with implications for mechanisms of naïve and immune repertoire selection.
Open AccessHeavy Chain CDR 3 and Junctional Length Biases in Human Antibody Repertoires Associated with Heavy and Light Chain Germline Utilization
Antibody variable domain sequence diversity is generated by recombination of germline segments. The third complementarity-determining region of the heavy chain (CDR H3) is the region of highest sequence diversity and is formed by the joining of heavy chain V H , D H and J H germline segments combined with random nucleotide trimming and additions between these segments. We show that CDR H3 length distribution is biased in human antibody repertoires as a function of V H , V L and J H germline segment utilization. Most length biases are apparent in the naïve B cell compartment, with a significant bias towards shorter CDR H3 sequences observed in association with a subset of V H and V L germlines in the antigen experienced compartment. Similar biases were not observed in nonproductive heavy chain recombination products, indicating selection of the repertoire during B cell maturation as a major driver of the length biases. Some V H -associated CDR H3 length biases are dependent on utilization of specific J H germline segments in a manner not directly linked to J H segment length in the germline, but are rather associated with selection of differentially trimmed J H segments in the naïve compartment. In addition, D H segment and N-region random nucleotide insertion lengths within CDR H3 in the naïve compartment were also biased by specific V H /J H germline combinations, indicating a complex set of constraints between germline segments selected during repertoire maturation. Our findings reveal biases in the antibody diversity landscape shaped by V H , V L , and J H germline features with implications for mechanisms of naïve and immune repertoire selection.
- Open Access
- 25 September 2018
- Published by Taylor & Francis in mAbs
- Vol. 10 (8) , 1281-1290
Prediction of methionine oxidation risk in monoclonal antibodies using a machine learning method
Monoclonal antibodies (mAbs) have become a major class of protein therapeutics that target a spectrum of diseases ranging from cancers to infectious diseases. Similar to any protein molecule, mAbs are susceptible to chemical modifications during the manufacturing process, long-term storage, and in vivo circulation that can impair their potency. One such modification is the oxidation of methionine residues. Chemical modifications that occur in the complementarity-determining regions (CDRs) of mAbs can lead to the abrogation of antigen binding and reduce the drug’s potency and efficacy. Thus, it is highly desirable to identify and eliminate any chemically unstable residues in the CDRs during the therapeutic antibody discovery process. To provide increased throughput over experimental methods, we extracted features from the mAbs’ sequences, structures, and dynamics, used random forests to identify important features and develop a quantitative and highly predictive in silico methionine oxidation model.
- 25 September 2018
- Published by Taylor & Francis in mAbs
- Vol. 10 (8) , 1281-1290
Open AccessPrediction of methionine oxidation risk in monoclonal antibodies using a machine learning method
Monoclonal antibodies (mAbs) have become a major class of protein therapeutics that target a spectrum of diseases ranging from cancers to infectious diseases. Similar to any protein molecule, mAbs are susceptible to chemical modifications during the manufacturing process, long-term storage, and in vivo circulation that can impair their potency. One such modification is the oxidation of methionine residues. Chemical modifications that occur in the complementarity-determining regions (CDRs) of mAbs can lead to the abrogation of antigen binding and reduce the drug’s potency and efficacy. Thus, it is highly desirable to identify and eliminate any chemically unstable residues in the CDRs during the therapeutic antibody discovery process. To provide increased throughput over experimental methods, we extracted features from the mAbs’ sequences, structures, and dynamics, used random forests to identify important features and develop a quantitative and highly predictive in silico methionine oxidation model.
- Open Access
- 2 April 2018
- Published by Elsevier in Journal of Immunological Methods
- Vol. 455, 88-94
Barcoded sequencing workflow for high throughput digitization of hybridoma antibody variable domain sequences
- 2 April 2018
- Published by Elsevier in Journal of Immunological Methods
- Vol. 455, 88-94
Open AccessBarcoded sequencing workflow for high throughput digitization of hybridoma antibody variable domain sequences
- 17 August 2017
- Published by Wiley in Biotechnology Journal
- Vol. 13 (1)
Supplementation of Nucleosides During Selection can Reduce Sequence Variant Levels in CHO Cells Using GS/MSX Selection System
In the process of generating stable monoclonal antibody (mAb) producing cell lines, reagents such as methotrexate (MTX) or methionine sulfoximine (MSX) are often used. However, using such selection reagent(s) increases the possibility of having higher occurrence of sequence variants in the expressed antibody molecules due to the effects of MTX or MSX on de novo nucleotide synthesis. Since MSX inhibits glutamine synthase (GS) and results in both amino acid and nucleoside starvation, it is questioned whether supplementing nucleosides into the media could lower sequence variant levels without affecting titer. The results show that the supplementation of nucleosides to the media during MSX selection decreased genomic DNA mutagenesis rates in the selected cells, probably by reducing nucleotide mis‐incorporation into the DNA. Furthermore, addition of nucleosides enhance clone recovery post selection and does not affect antibody expression. It is further observed that nucleoside supplements lowered DNA mutagenesis rates only at the initial stage of the clone selection and do not have any effect on DNA mutagenesis rates after stable cell lines are established. Therefore, the data suggests that addition of nucleosides during early stages of MSX selection can lower sequence variant levels without affecting titer or clone stability in antibody expression.
- 17 August 2017
- Published by Wiley in Biotechnology Journal
- Vol. 13 (1)
Supplementation of Nucleosides During Selection can Reduce Sequence Variant Levels in CHO Cells Using GS/MSX Selection System
In the process of generating stable monoclonal antibody (mAb) producing cell lines, reagents such as methotrexate (MTX) or methionine sulfoximine (MSX) are often used. However, using such selection reagent(s) increases the possibility of having higher occurrence of sequence variants in the expressed antibody molecules due to the effects of MTX or MSX on de novo nucleotide synthesis. Since MSX inhibits glutamine synthase (GS) and results in both amino acid and nucleoside starvation, it is questioned whether supplementing nucleosides into the media could lower sequence variant levels without affecting titer. The results show that the supplementation of nucleosides to the media during MSX selection decreased genomic DNA mutagenesis rates in the selected cells, probably by reducing nucleotide mis‐incorporation into the DNA. Furthermore, addition of nucleosides enhance clone recovery post selection and does not affect antibody expression. It is further observed that nucleoside supplements lowered DNA mutagenesis rates only at the initial stage of the clone selection and do not have any effect on DNA mutagenesis rates after stable cell lines are established. Therefore, the data suggests that addition of nucleosides during early stages of MSX selection can lower sequence variant levels without affecting titer or clone stability in antibody expression.
- 3 January 2017
- Published by American Chemical Society (ACS) in ACS Chemical Biology
- Vol. 12 (2) , 510-518
Profiling Protease Specificity: Combining Yeast ER Sequestration Screening (YESS) with Next Generation Sequencing
An enzyme engineering technology involving yeast endoplasmic reticulum (ER) sequestration screening (YESS) has been recently developed. Here, a new method is established, in which the YESS platform is combined with NextGen sequencing (NGS) to enable a comprehensive survey of protease specificity. In this approach, a combinatorial substrate library is targeted to the yeast ER and transported through the secretory pathway, interacting with any protease(s) residing in the ER. Multicolor FACS screening is used to isolate cells labeled with fluorophore-conjugated antibodies, followed by NGS to profile the cleaved substrates. The YESS-NGS method was successfully applied to profile the sequence specificity of the wild-type and an engineered variant of the tobacco etch mosaic virus protease. Proteolysis in the yeast secretory pathway was also mapped for the first time in vivo revealing a major cleavage pattern of Ali/Leu-X-Lys/Arg-Arg. Here Ali is any small aliphatic residue, but especially Leu. This pattern was verified to be due to the well-known endogenous protease Kex2 after comparison to a newly generated Kex2 knockout strain as well as cleavage of peptides with recombinant Kex2 in vitro. This information is particularly important for those using yeast display technology, as library members with Ali/Leu-X-Lys/Arg-Arg patterns are likely being removed from screens via Kex2 cleavage without the researcher’s knowledge.
- 3 January 2017
- Published by American Chemical Society (ACS) in ACS Chemical Biology
- Vol. 12 (2) , 510-518
Profiling Protease Specificity: Combining Yeast ER Sequestration Screening (YESS) with Next Generation Sequencing
An enzyme engineering technology involving yeast endoplasmic reticulum (ER) sequestration screening (YESS) has been recently developed. Here, a new method is established, in which the YESS platform is combined with NextGen sequencing (NGS) to enable a comprehensive survey of protease specificity. In this approach, a combinatorial substrate library is targeted to the yeast ER and transported through the secretory pathway, interacting with any protease(s) residing in the ER. Multicolor FACS screening is used to isolate cells labeled with fluorophore-conjugated antibodies, followed by NGS to profile the cleaved substrates. The YESS-NGS method was successfully applied to profile the sequence specificity of the wild-type and an engineered variant of the tobacco etch mosaic virus protease. Proteolysis in the yeast secretory pathway was also mapped for the first time in vivo revealing a major cleavage pattern of Ali/Leu-X-Lys/Arg-Arg. Here Ali is any small aliphatic residue, but especially Leu. This pattern was verified to be due to the well-known endogenous protease Kex2 after comparison to a newly generated Kex2 knockout strain as well as cleavage of peptides with recombinant Kex2 in vitro. This information is particularly important for those using yeast display technology, as library members with Ali/Leu-X-Lys/Arg-Arg patterns are likely being removed from screens via Kex2 cleavage without the researcher’s knowledge.