Abstract
This project brings together two comparatively-novel approaches – antibody repertoire sequencing using NGS (Rep-seq), and deep learning – to help characterise lymph node antibody repertoires. Rep-seq datasets commonly comprise millions of antibody sequences – too large for most conventional analytic approaches. Recently, deep learning techniques (e.g. convolution neural networks) have been applied to large, high-dimensional sequence datasets.
The aim of this project is to create deep learning predictors for identifying key attributes of antibody repertoires that have been extracted from uninvolved (non-tumour bearing) lymph nodes.
References
1. Irshad, Sheeba, et al. “RORγt+ Innate Lymphoid Cells Promote Lymph Node Metastasis of Breast Cancers.” Cancer research 77.5 (2017): 1083-1096.
2. Lees, William D., and Adrian J. Shepherd. “Utilities for high-throughput analysis of B-cell clonal lineages.” Journal of immunology research 2015 (2015).
3. Zeng, Haoyang, et al. “Convolutional neural network architectures for predicting DNA–protein binding.” Bioinformatics 32.12 (2016): i121-i127.
4. Shepherd, Adrian J. Second-Order Methods for Neural Networks: Fast and Reliable Training Methods for Multi-Layer Perceptrons. Perspectives in Neural Computing series, Springer, May 1997.
5. DeKosky, Brandon J., et al. “In-depth determination and analysis of the human paired heavy-and light-chain antibody repertoire.” Nature medicine 21.1 (2015): 86-91.