BKBenjamin P. Kellman

Benjamin P. Kellman

*This is an algorithmically generated author record based on publicly available information.

Current Organizations / Institutes

The organizations/institutes from the scholar's latest article

Previous Organizations / Institutes

The organizations / institutes from the scholar's other articles in the past ten years


Benjamin P. Kellman has a total of 290 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • Open Access

    mRNA-Protein Coordination is Contextualized by Metastatic Biological Phenotypes

    by Ronit Sharma, Nikolaos Meimetis, Arjana Begzati, Shashwat Depali Nagar, Benjamin Kellman, Hratch M. Baghdassarian

    A central goal of conducting omics measurements is to understand how molecular features inform higher-order cell- and tissue-level phenotypes. In particular, multi-omics offers insights into how information encoded by the genome is coordinated through biological layers, resulting in functional outputs 1 . Due to myriad post-transcriptional regulatory processes, the coordination between mRNA and protein cannot be simply reduced to gene-wise correlation. Yet, both modalities have been shown to serve as representations of biological state, and multi-omics integration has been used to improve these representations. Multi-omics approaches typically do not focus on how mRNA and protein features coordinate, but rather use the additional information for improved prediction or feature selection. Here, instead, we showed that standard linear machine learning models provide an understanding of transcriptomic and proteomic coordination in the context of a biological phenotype of interest, in this case cancer metastasis. We find that, in the context of metastasis, a select subset of proteomic features—reflecting a more concentrated signal relative to the broadly distributed transcriptomic signal—offers additional information to that encoded by transcriptomics, as demonstrated by improved model performance when integrating the two modalities and the relative feature importance of proteomics. Top features show a depletion of gene-product overlap across modalities, indicating that the model primarily leverages instances in which the two modalities are providing complementary information with respect to phenotype. However, in instances when both modalities are selected for a given gene product, there is high information consistency that synergistically bolsters phenotype prediction. Altogether, by using model fits that relate both modalities to phenotype, we observe a nuanced coordination of protein and mRNA, in which both modalities tend to provide consistent information about phenotype, yet benefits remain to incorporating a combination of both complementary and reinforcing signals across modalities.

  • Open Access

    Protein structure, a genetic encoding for glycosylation

    by Ben Kellman, Daniel Sandoval, Olga Zaytseva, Kelly Brock, Sabyasachi Baboo, Daniela Nachmanson, Edward Irvine, Erick Armingol, Nathan Mih, Yujie Zhang, et al.

    DNA, RNA, and proteins are synthesized using template molecules, but glycosylation is not believed to be constrained by a template. However, if cellular environment is the only determinant of glycosylation, all sites should receive the same glycans on average. This template-free assertion is inconsistent with observations of microheterogeneity-wherein each site receives distinct and reproducible glycan structures. Here, we test the assumption of template-free glycan biosynthesis. Through structural analysis of site-specific glycosylation data, we find protein-sequence and structural features that predict specific glycan features. To quantify these relationships, we present a new amino acid substitution matrix that describes glycoimpact-how glycosylation varies with protein structure. High-glycoimpact amino acids co-evolve with glycosites, and glycoimpact is high when estimates of amino acid conservation and variant pathogenicity diverge. We report hundreds of disease variants near glycosites with high-glycoimpact, including several with known links to aberrant glycosylation (e.g., Oculocutaneous Albinism, Jakob-Creutzfeldt disease, Gerstmann-Straussler-Scheinker, and Gaucher's Disease). Finally, we validate glycoimpact quantification by studying oligomannose-complex glycan ratios on HIV ENV, differential sialylation on IgG3 Fc, differential glycosylation on SARS-CoV-2 Spike, and fucose-modulated function of a tuberculosis monoclonal antibody. In all, we show glycan biosynthesis is accurately guided by specific, genetically-encoded rules, and this presents a plausible refutation to the assumption of template-free glycosylation.


  • Open Access

    Decoding glycosylation potential from protein structure across human glycoproteins with a multi-view recurrent neural network

    by Benjamin P. Kellman, Julien Mariethoz, Yujie Zhang, Sigal Shaul, Mia Alteri, Daniel Sandoval, Mia Jeffris, Erick Armingol, Bokan Bao, Frederique Lisacek, et al.

    Glycosylation is described as a non-templated biosynthesis. Yet, the template-free premise is antithetical to the observation that different N-glycans are consistently placed at specific sites. It has been proposed that glycosite-proximal protein structures could constrain glycosylation and explain the observed microheterogeneity. Using site-specific glycosylation data, we trained a hybrid neural network to parse glycosites (recurrent neural network) and match them to feasible N-glycosylation events (graph neural network). From glycosite-flanking sequences, the algorithm predicts most human N-glycosylation events documented in the GlyConnect database and proposed structures corresponding to observed monosaccharide composition of the glycans at these sites. The algorithm also recapitulated glycosylation in Enhanced Aromatic Sequons, SARS-CoV-2 spike, and IgG3 variants, thus demonstrating the ability of the algorithm to predict both glycan structure and abundance. Thus, protein structure constrains glycosylation, and the neural network enables predictive in silico glycosylation of uncharacterized or novel protein sequences and genetic variants.


  • Open Access

    Boosting with adjuvanted SCB-2019 elicits superior Fcγ-receptor engagement driven by IgG3 to SARS-CoV-2 spike

    by Wonyeong Jung, Dansu Yuan, Benjamin Kellman, Isabela Garrido da Silva Gonzalez, Ralf Clemens, Eveline Pipolo Milan, Eduardo Sprinz, José Cerbino Neto, Igor Smolenov, Galit Alter, et al.

    With the continued emergence of variants of concern, the global threat of COVID-19 persists, particularly in low- and middle-income countries with limited vaccine access. Protein-based vaccines, such as SCB-2019, can be produced on a large scale at a low cost while antigen design and adjuvant use can modulate efficacy and safety. While effective humoral immunity against SARS-CoV-2 variants has been shown to depend on both neutralization and Fc-mediated immunity, data on the effectiveness of protein-based vaccines with enhanced Fc-mediated immunity is limited. Here, we assess the humoral profile, including antibody isotypes, subclasses, and Fc receptor binding generated by a boosting with a recombinant trimer-tag protein vaccine SCB-2019. Individuals who were primed with 2 doses of the ChAdOx1 vaccine were equally divided into 4 groups and boosted with following formulations: Group 1: 9 μg SCB-2019 and Alhydrogel; Group 2: 9 μg SCB-2019, CpG 1018, and Alhydrogel; Group 3: 30 μg SCB-2019, CpG 1018, and Alhydrogel; Group 4: ChAdOx1. Group 3 showed enhanced antibody FcγR binding against wild-type and variants compared to Groups 1 and 2, showing a dose-dependent enhancement of immunity conferred by the SCB-2019 vaccine. Moreover, from day 15 after vaccination, Group 3 exhibited higher IgG3 and FcγR binding across variants of concerns, including Omicron and its subvariants, compared to the ChAdOx1-boosted individuals. Overall, this highlights the potential of SCB-2019 as a cost-efficient boosting regimen effective across variants of concerns.


  • Humoral immunity to an endemic coronavirus is associated with postacute sequelae of COVID-19 in individuals with rheumatic diseases

    by Jonathan D. Herman, Caroline Atyeo, Yonatan Zur, Claire E. Cook, Naomi J. Patel, Kathleen M. Vanni, Emily N. Kowalski, Grace Qian, Shruthi Srivatsan, Nancy A. Shadick, et al.

    Beyond the acute illness caused by severe acute respiratory coronavirus 2 (SARS-CoV-2) infection, about one-fifth of infections result in long-term persistence of symptoms despite the apparent clearance of infection. Insights into the mechanisms that underlie postacute sequelae of COVID-19 (PASC) will be critical for the prevention and clinical management of long-term complications of COVID-19. Several hypotheses have been proposed that may account for the development of PASC, including persistence of virus and dysregulation of immune responses. Among the immunological changes noted in PASC, alterations in humoral immunity have been observed in some patient subsets. To begin to determine whether SARS-CoV-2– or other pathogen-specific humoral immune responses evolve uniquely in PASC, we performed comprehensive antibody profiling against SARS-CoV-2, a panel of endemic pathogens, and a panel of routine vaccine antigens using systems serology in two cohorts of patients with preexisting systemic autoimmune rheumatic disease (SARD) who either developed or did not develop PASC. A distinct qualitative shift observed in Fcγ receptor (FcγR) binding was observed in individuals with PASC. Specifically, individuals with PASC harbored weaker FcγR-binding anti–SARS-CoV-2 antibodies and stronger FcγR-binding antibody responses against the endemic coronavirus OC43. Individuals with PASC developed an OC43 S2-specific antibody response with stronger FcγR binding, linked to cross-reactivity across SARS-CoV-2 and common coronaviruses. These findings identify previous coronavirus imprinting as a potential marker for the development of PASC in individuals with SARDs.


  • Open Access

    Preparing glycomics data for robust statistical analysis with GlyCompareCT

    by Yujie Zhang, Sridevi Krishnan, Bokan Bao, Austin W.T. Chiang, James T. Sorrentino, Song-Min Schinn, Benjamin P. Kellman, Nathan E. Lewis


  • Simple and practical sialoglycan encoding system reveals vast diversity in nature and identifies a universal sialoglycan-recognizing probe derived from AB5 toxin B subunits

    by Aniruddha Sasmal, Naazneen Khan, Zahra Khedri, Benjamin P Kellman, Saurabh Srivastava, Andrea Verhagen, Hai Yu, Anders Bech Bruntse, Sandra Diaz, Nissi Varki, et al.

    Vertebrate sialic acids (Sias) display much diversity in modifications, linkages, and underlying glycans. Slide microarrays allow high-throughput explorations of sialoglycan–protein interactions. A microarray presenting ~150 structurally defined sialyltrisaccharides with various Sias linkages and modifications still poses challenges in planning, data sorting, visualization, and analysis. To address these issues, we devised a simple 9-digit code for sialyltrisaccharides with terminal Sias and underlying two monosaccharides assigned from the nonreducing end, with 3 digits assigning a monosaccharide, its modifications, and linkage. Calculations based on the encoding system reveal >113,000 likely linear sialyltrisaccharides in nature. Notably, a biantennary N-glycan with 2 terminal sialyltrisaccharides could thus have >1010 potential combinations and a triantennary N-glycan with 3 terminal sequences, >1015 potential combinations. While all possibilities likely do not exist in nature, sialoglycans encode enormous diversity. While glycomic approaches are used to probe such diverse sialomes, naturally occurring bacterial AB5 toxin B subunits are simpler tools to track the dynamic sialome in biological systems. Sialoglycan microarray was utilized to compare sialoglycan-recognizing bacterial toxin B subunits. Unlike the poor correlation between B subunits and species phylogeny, there is stronger correlation with Sia-epitope preferences. Further supporting this pattern, we report a B subunit (YenB) from Yersinia enterocolitica (broad host range) recognizing almost all sialoglycans in the microarray, including 4-O-acetylated-Sias not recognized by a Yersinia pestis orthologue (YpeB). Differential Sia-binding patterns were also observed with phylogenetically related B subunits from Escherichia coli (SubB), Salmonella Typhi (PltB), Salmonella Typhimurium (ArtB), extra-intestinal E.coli (EcPltB), Vibrio cholera (CtxB), and cholera family homologue of E. coli (EcxB).


  • Open Access

    Preparing glycomics data for robust statistical analysis with GlyCompareCT

    by Yujie Zhang, Sridevi Krishnan, Bokan Bao, Austin W.T. Chiang, James T. Sorrentino, Song-Min Schinn, Benjamin P. Kellman, Nathan E. Lewis

    Summary: Glycomics data are rapidly increasing in scale and diversity. Complexities in glycan biosynthesis (hierarchy, competition, and compartmentalization) make preprocessing critical to address resulting sparsity (many similar glycosylation profiles may share few common glycans) and non-independence (substrate-competition in glycan biosynthesis results in non-independence incompatible with many statistical methods). Here, we present GlyCompareCT, a portable command-line tool, to address these challenges thereby facilitating downstream analyses. Given glycan abundances, GlyCompareCT conducts substructure decomposition to quantify hidden biosynthetic intermediate abundance and relationships between measured glycans. Thus, GlyComparCT mitigates sparsity and makes interdependence explicit thereby increasing statistical power. Ultimately, GlyComparCT is a user-friendly implementation of substructure analysis designed to increase accessibility, interoperability, and scope and consistency in glycomics analysis. Availability and implementation: Source code, test data, and instructions for GlyCompareCT v1.1.0 are available at: https://github.com/LewisLabUCSD/GlyCompareCT Supplementary information: https://github.com/LewisLabUCSD/GlyCompareCT/raw/main/Supplementary%20Material.pdf

  • Open Access

    Multiplex genome editing of mammalian cells for producing recombinant heparin

    by Bryan E. Thacker, Kristen J. Thorne, Colin Cartwright, Jeeyoung Park, Kimberly Glass, Annie Chea, Benjamin P. Kellman, Nathan E. Lewis, Zhenping Wang, Anna Di Nardo, et al.


  • Open Access

    Correcting for sparsity and interdependence in glycomics by accounting for glycan biosynthesis

    by Bokan Bao, Benjamin P. Kellman, Austin W. T. Chiang, Yujie Zhang, James T. Sorrentino, Austin K. York, Mahmoud A. Mohammad, Morey W. Haymond, Lars Bode, Nathan E. Lewis

    Glycans are fundamental cellular building blocks, involved in many organismal functions. Advances in glycomics are elucidating the essential roles of glycans. Still, it remains challenging to properly analyze large glycomics datasets, since the abundance of each glycan is dependent on many other glycans that share many intermediate biosynthetic steps. Furthermore, the overlap of measured glycans can be low across samples. We address these challenges with GlyCompare, a glycomic data analysis approach that accounts for shared biosynthetic steps for all measured glycans to correct for sparsity and non-independence in glycomics, which enables direct comparison of different glycoprofiles and increases statistical power. Using GlyCompare, we study diverse N-glycan profiles from glycoengineered erythropoietin. We obtain biologically meaningful clustering of mutant cell glycoprofiles and identify knockout-specific effects of fucosyltransferase mutants on tetra-antennary structures. We further analyze human milk oligosaccharide profiles and find mother’s fucosyltransferase-dependent secretor-status indirectly impact the sialylation. Finally, we apply our method on mucin-type O-glycans, gangliosides, and site-specific compositional glycosylation data to reveal tissues and disease-specific glycan presentations. Our substructure-oriented approach will enable researchers to take full advantage of the growing power and size of glycomics data.