Blending deep learning with probabilistic mechanistic models to predict and understand the evolution and function of adaptive immune receptors

将深度学习与概率机制模型相结合，以预测和理解适应性免疫受体的进化和功能

• 批准号: 10415985
• 负责人: Frederick Albert Matsen
• 金额: $ 68.96万
• 依托单位: FRED HUTCHINSON CANCER CENTER
• 依托单位国家: 美国
• 财政年份: 2019
• 资助国家: 美国
• 起止时间: 2019-06-01 至 2025-05-31
• 项目状态: 未结题
• 来源: https://reporter.nih.gov/project-details/10415985
• 关键词: 3-Dimensional; Algorithms; Animal Model; Antibodies; Antibody Affinity; Antigens; Architecture; Autoimmune Diseases; Automobile Driving; Big Data; Binding; Biochemical; Biochemical Process; Categories; Cell Line; Characteristics; Collection; Complement; Complex; Data; Data Set; Dependence; Diagnosis; Disease; Entropy; Evolution; Exposure to; Foundations; Gene Conversion; Generations; Goals; High-Throughput DNA Sequencing; High-Throughput Nucleotide Sequencing; Human; Hybrids; Immune; Immune response; Immune system; Immunoglobulin Somatic Hypermutation; Immunologic Memory; Immunologic Receptors; Immunological Models; Immunologics; Immunologist; Immunology; Immunotherapy; In Vitro; Individual; Knock-out; Knowledge; Laboratories; Machine Learning; Medical; Methods; Modeling; Modification; Mutation; Pathway interactions; Population; Procedures; Process; Property; Prophylactic treatment; Resolution; Sampling; Science; Statistical Distributions; Statistical Methods; Statistical Models; T-Cell Receptor; T-Lymphocyte; T-cell receptor repertoire; Techniques; Technology; Testing; Time; Training; Update; V(D)J Recombination; Vaccination; Vaccine Design; Vaccines; Validation; Work; algorithm training; analytical tool; base; biochemical model; cancer immunotherapy; cancer therapy; complex data; data complexity; deep learning; deep learning model; deep neural network; deep sequencing; design; experimental study; fighting; functional group; in vivo; insertion/deletion mutation; large datasets; machine learning method; markov model; pathogen; progenitor; receptor; repaired; response; success; tool

Project Summary Scientific understanding of adaptive immune receptors (i.e. antibodies and T cell receptors) has the potential to revolutionize prophylaxis, diagnosis, and treatment of disease. High‐throughput DNA sequencing and functional experiments have now brought the study of adaptive immune receptors into the big‐data era. To realize this potential of these data they must be matched with appropriately powerful analytical techniques. Existing probabilistic and mechanistic models are insufficient to capture the complexities of these data, while a naïve application of machine learning cannot leverage our profound existing knowledge of the immune system. The goal of this project is to blend deep learning with mechanistic modeling in order to predict and understand the evolution and function of adaptive immune receptors. Aim 1: Develop generative models of immune receptor sequences that capture the complexity of real adaptive immune receptor repertoires. These will combine deep learning along with our knowledge of VDJ recombination, and provide a rigorous platform for detailed repertoire comparison. Aim 2: Develop quantitative mechanistic models of antibody somatic hypermutation that incorporate the underlying biochemical processes. Estimate intractable likelihoods using deep learning to infer important latent variables, and validate models using knock‐out experiments in cell lines. Aim 3: Develop hybrid deep learning models to predict binding properties from sequence data, combining large experimentally‐derived binding data with even larger sets of immune sequences from human immune memory samples. Incorporate structural information via 3D convolution or distance‐based penalties. These tools will reveal the full power of immune repertoire data for medical applications. We will obtain more rigorous comparisons of repertoires via their distribution in a relevant space. These will reveal the effects of immune perturbations such as vaccination and disease, allowing us to pick out sequences that are impacted by these perturbations. We will have a greater quantitative understanding of somatic hypermutation in vivo, and statistical models that appropriately capture long‐range effects of collections of mutations. We will also have algorithms that will be able to combine repertoire data and sparse binding data to predict binding properties. Put together, these advances will enable rational vaccine design, treatment for autoimmune disease, and identification of T cells that are promising candidates for cancer immunotherapy.

项目总结

项目成果

Inference of B cell clonal families using heavy/light chain pairing information.

• DOI: 10.1371/journal.pcbi.1010723
• 发表时间: 2022-11
• 期刊: PLoS computational biology
• 影响因子: 4.3

Enabling Inference for Context-Dependent Models of Mutation by Bounding the Propagation of Dependency.

通过限制依赖性的传播来实现上下文相关的突变模型的推理。

• DOI: 10.1089/cmb.2021.0644
• 发表时间: 2022
• 期刊: Journal of computational biology : a journal of computational molecular cell biology
• 作者: Matsen4th,FrederickA;Ralph,PeterL
• 通讯作者: Ralph,PeterL

Reconstruction of a polyclonal ADCC antibody repertoire from an HIV-1 non-transmitting mother.

• DOI: 10.1016/j.isci.2023.106762
• 发表时间: 2023-05-19
• 期刊: ISCIENCE
• 影响因子: 5.8
• 作者: Yaffe, Zak A.;Ding, Shilei;Sung, Kevin;Chohan, Vrasha;Marchitto, Lorie;Doepker, Laura;Ralph, Duncan;Nduati, Ruth;Matsen, Frederick A.;Finzi, Andres;Overbaugh, Julie
• 通讯作者: Overbaugh, Julie

Lack of Evidence for a Substantial Rate of Templated Mutagenesis in B Cell Diversification.

• DOI: 10.4049/jimmunol.2000092
• 发表时间: 2020-08-15
• 期刊: Journal of immunology (Baltimore, Md. : 1950)
• 作者: Fukuyama J;Olson BJ;Matsen FA 4th
• 通讯作者: Matsen FA 4th