Key Parameters of Tumor Epitope Immunogenicity Revealed Through a Consortium Approach Improve Neoantigen Prediction

Wells, Daniel K.; Buuren, Marit M. van; Dang, Kristen K.; Hubbard-Lucey, Vanessa M.; Sheehan, Kathleen C. F.; Campbell, Katie M.; Lamb, Andrew; Ward, Jeffrey P.; Sidney, John; Blázquez, Ana B.; Rech, Andrew J.; Zaretsky, Jesse M.; Comin-Anduix, Begoñya; Ng, Alphonsus H. C.; Chour, William; Yu, Thomas; Rizvi, Hira; Chen, Jia M.; Manning, Patrice; Steiner, Gabriela; Doan, Xengie; Khan, Aly A.; Lugade, Amit A.; Lazic, Ana Mijalkovic; Frentzen, Angela; Tadmor, Arbel D.; Sasson, Ariella; Av, Rao; Pei, Baikang; Schrörs, Barbara; Berent-Maoz, Beata; Carreño, Beatriz M.; Song, Bin; Peters, Bjoern; Li, Bo; Higgs, Brandon W.; Stevenson, Brian J.; Iseli, Christian; Miller, Christopher A.; Morehouse, Christopher; Melief, Cornelis J.M.; Puig-Saus, Cristina; Beek, Daphne van; Balli, David; Gfeller, David; Haussler, David; Jäger, Dirk; Cortes, Eduardo; Esaulova, Ekaterina; Sherafat, Elham; Arcila, Francisco; Bartha, Gábor; Li, Geng; Coukos, George; Richard, Guilhem; Han, Chang; Si, Han; Zörnig, Inka; Xénarios, Ioannis; Măndoiu, Ion; Kooi, Irsan; Conway, James; Kessler, Jan H.; Greenbaum, Jason A.; Perera, Jason; Harris, Jason; Hundal, Jasreet; Shelton, Jennifer; Wang, Jianmin; Wang, Hua; Greshock, Joel; Blake, Jonathon; Szustakowski, Joseph D.; Kodysh, Julia; Forman, Juliet; Wei, Lei; Lee, Leo J.; Fanchi, Lorenzo F.; Slagter, Maarten; Lang, Maren; Mueller, Markus; Löwer, Martin; Vormehr, Mathias; Artyomov, Maxim N.; Kuziora, Michael; Princiotta, Michael F.; Bassani-Sternberg, Michal; Macabali, Mignonette; Kojicic, Milica R.; Yang, Naibo; Raicevic, Nevena M. Ilic; Guex, Nicolas; Robine, Nicolas; Halama, Niels; Skundric, Nikola; Milićević, Ognjen; Gellert, Pascal; Jongeneel, Patrick; Charoentong, Pornpimol; Srivastava, Pramod K.; Tanden, Prateek; Shah, Priyanka; Hu, Qiang; Gupta, Ravi; Chen, Richard; Petit, Robert G.; Ziman, Robert; Hilker, Rolf; Shukla, Sachet A.; Seesi, Sahar Al; Boyle, Sean M.; Qiu, Si; Sarkizova, Siranush; Salama, Sofie R.; Liu, Song; Wu, Song; Sridhar, Srinivas; Ketelaars, Steven L. C.; Jhunjhunwala, Suchit; Shcheglova, Tatiana; Schuepbach, Thierry; Creasy, Todd; Josipovic, Veliborka; Kovačević, Vladimir; Fu, Weixuan; Krebber, Willem-Jan; Hsu, Yi‐Hsiang; Sebastian, Yinong; Yalcin, Zeynep Kosaloglu; Huang, Zhiqin; Merghoub, Taha; Guinney, Justin; Kolom, Adam; Selinsky, Cheryl; Ribas, Antoni; Hellmann, Matthew D.; Hacohen, Nir; Sette, Alessandro; Heath, James R.; Bhardwaj, Nina; Ramsdell, Fred; Schreiber, Robert D.; Schumacher, Ton N.; Kvistborg, Pia; Defranoux, Nadine

doi:10.1016/j.cell.2020.09.015

Cited by 371 publications

(468 citation statements)

References 80 publications

Supporting

Mentioning

455

Contrasting

Order By: Relevance

“…For the tumor neoantigen test dataset, we considered: a restricted dataset of the (1) top 20 or (2) top 50 immunogenic peptides predictions for each algorithm’s or (3) overall sensitivity. The top 20 or 50 immunogenic peptides were purposely selected as these are the same number of peptides considered in prior related discovery or clinical reports [20]. For the sensitivity analysis, a threshold of 0.5 was used for DeepImmuno-CNN and DeepHLApan and a threshold of 0 for the IEDB default classification algorithm, which has a distinct scoring range.…”

Section: Resultsmentioning

confidence: 99%

“…At the end of cross validation, the scores for each evaluation metric were averaged over the ten testing subsets as the model’s performance. We selected two independent test datasets for further evaluation: 1) 637 experimentally tested tumor specific neoantigens from the Tumor Neoantigen Selection Alliance (TESLA) [20], and 2) 100 SARS-Cov-2 peptides [1,20] tested for their immunogenicity in convalescent and unexposed subjects, respectively.…”

Section: Methodsmentioning

confidence: 99%

See 1 more Smart Citation

DeepImmuno: Deep learning-empowered prediction and generation of immunogenic peptides for T cell immunity

Iyer

Prasath

et al. 2020

Preprint

View full text Add to dashboard Cite

T-cells play an essential role in the adaptive immune system by seeking out, binding and destroying foreign antigens presented on the cell surface of diseased cells. An improved understanding of T-cell immunity will greatly aid in the development of new cancer immunotherapies and vaccines for life threatening pathogens. Central to the design of such targeted therapies are computational methods to predict non-native epitopes to elicit a T cell response, however, we currently lack accurate immunogenicity inference methods. Another challenge is the ability to accurately simulate immunogenic peptides for specific human leukocyte antigen (HLA) alleles, for both synthetic biological applications and to augment real training datasets. Here, we proposed a beta-binomial distribution approach to derive epitope immunogenic potential from sequence alone. We conducted systematic benchmarking of five traditional machine learning (ElasticNet, KNN, SVM, Random Forest, AdaBoost) and three deep learning models (CNN, ResNet, GNN) using three independent prior validated immunogenic peptide collections (dengue virus, cancer neoantigen and SARS-Cov-2). We chose the CNN model as the best prediction model based on its adaptivity for small and large datasets, and performance relative to existing methods. In addition to outperforming two highly used immunogenicity prediction algorithms, DeepHLApan and IEDB, DeepImmuno-CNN further correctly predicts which residues are most important for T cell antigen recognition. Our independent generative adversarial network (GAN) approach, DeepImmuno-GAN, was further able to accurately simulate immunogenic peptides with physiochemical properties and immunogenicity predictions similar to that of real antigens. We provide DeepImmuno-CNN as source code and an easy-to-use web interface.Data AvailabilityDeepImmuno Python3 code is available at https://github.com/frankligy/DeepImmuno. The DeepImmuno web portal is available from https://deepimmuno.herokuapp.com. The data in this article is available in GitHub and supplementary materials.

show abstract

Section: Resultsmentioning

confidence: 99%

Section: Methodsmentioning

confidence: 99%

DeepImmuno: Deep learning-empowered prediction and generation of immunogenic peptides for T cell immunity

Iyer

Prasath

et al. 2020

Preprint

View full text Add to dashboard Cite

show abstract

“…Recently, the Tumor Neoantigen Selection Alliance (TESLA), a group of 25 teams that independently predicted and ranked neoantigens from a common data set, published their findings on features important for neoantigen prediction 17 . They made the unexpected observation that among the 37 positively validated neoantigen candidates, none of the peptides had a mutation at position 2, a common anchor position for a range of HLA alleles, despite a high number of prioritized neoantigens with a position 2 mutation.…”

Section: Discussionmentioning

confidence: 99%

Computational prediction of MHC anchor locations guide neoantigen identification and prioritization

Xia¹,

McMichael²,

Supabphol

et al. 2020

Preprint

Self Cite

View full text Add to dashboard Cite

Neoantigens are novel peptide sequences resulting from somatic mutations in tumors that upon loading onto major histocompatibility complex (MHC) molecules allow recognition by T cells. Accurate neoantigen identification is thus critical for designing cancer vaccines and predicting response to immunotherapies. Neoantigen identification and prioritization relies on correctly predicting whether the presenting peptide sequence can successfully induce an immune response. As the majority of somatic mutations are SNVs, changes between wildtype and mutant peptide are subtle and require cautious interpretation. An important yet potentially underappreciated variable in neoantigen-prediction pipelines is the mutation position within the peptide relative to its anchor positions for the patient’s specific HLA alleles. While a subset of peptide positions is presented to the T-cell receptor for recognition, others are responsible for anchoring to the MHC, making these positional considerations critical for predicting T-cell responses. We computationally predicted high probability anchor positions for different peptide lengths for over 300 common HLA alleles and identified unique anchoring patterns among them. Analysis of 923 tumor samples shows that 7-41% of neoantigen candidates are potentially misclassified and can be rescued using allele-specific knowledge of anchor positions.

show abstract

“… 21 A delicate balancing between sensitivity and specificity is required in the pipeline for the prediction of candidates as it was substantiated by the comparative study of the TESLA consortium. 22 While filtering out 98% of non-immunogenic peptides with more than 70% precision is a leap forward, indications with low mutational burden which were not studied might require an even higher precision.…”

Section: Neoantigen-reactive T Cells Recognize Non-self Tumor-specifimentioning

confidence: 99%

Advances in identification and selection of personalized neoantigen/T-cell pairs for autologous adoptive T cell therapies

et al. 2021

View full text Add to dashboard Cite

Based on the success of tumor-infiltrating lymphocytes (TIL)-based therapies, personalized adoptive cell therapies (ACT) targeting neoantigens have the potential to become a disruptive technology and lead to highly effective treatments for cancer patients for whom no other options exist. ACT of TIL, peripheral blood or gene-engineered peripheral blood lymphocytes (PBLs) targeting neoantigens is a highly personalized intervention that requires three discrete steps: i) Identification of suitable personal targets (neoantigens), ii) selection of T cells or their T cell receptors (TCRs) that are specific for the identified neoantigens and iii) expansion of the selected T cell population or generation of sufficient number of TCR modified T cells. In this review, we provide an introduction into challenges and approaches to identify neoantigens and to select the Adoptive Cell Therapy, ACT, Neoantigen, T cell, Cancer respective neoantigen-reactive T cells for use in ACT.

show abstract

Key Parameters of Tumor Epitope Immunogenicity Revealed Through a Consortium Approach Improve Neoantigen Prediction

Cited by 371 publications

References 80 publications

DeepImmuno: Deep learning-empowered prediction and generation of immunogenic peptides for T cell immunity

DeepImmuno: Deep learning-empowered prediction and generation of immunogenic peptides for T cell immunity

Computational prediction of MHC anchor locations guide neoantigen identification and prioritization

Advances in identification and selection of personalized neoantigen/T-cell pairs for autologous adoptive T cell therapies

Contact Info

Product

Resources

About