Speaker
Description
RNA sequencing is a vital technology for immune repertoire profiling, allowing us to understand how the human adaptive immune system is structured. This technique enables clonal lineages to be tracked spatially and longitudinally: important for quantifying immune responses and the impact of immunotherapies \cite{oakesQuantitativeCharacterizationCell2017}. However, non-Poissonian sources of noise exist in the data that can make decomposing biological and technical signals challenging \cite{gierlinskiStatisticalModelsRNAseq2015}. Here, we conduct a variance decomposition experiment with blood draws from healthy volunteers, creating replicates at the repertoire sampling, library preparation, and sequencing levels. This strategy allows us to identify the contributions to noise from biological variation, molecular biology techniques, and short-read sequencing technologies. We then build an interpretable hierarchical statistical model informed by biological sampling mechanisms which captures the variance behaviour observed. We use this model to simulate the impact that sample properties, such as cell count, have upon the molecular mechanisms of noise. Finally, we use our mechanistic understanding to investigate suitable methods to account for non-Poissonian mean-variance relationships in adaptive immune repertoire sequencing data.
Bibliography
@article{oakesQuantitativeCharacterizationCell2017,
title = {Quantitative {{Characterization}} of the {{T Cell Receptor Repertoire}} of {{Na\"ive}} and {{Memory Subsets Using}} an {{Integrated Experimental}} and {{Computational Pipeline Which Is Robust}}, {{Economical}}, and {{Versatile}}},
author = {Oakes, Theres and Heather, James M. and Best, Katharine and {Byng-Maddick}, Rachel and Husovsky, Connor and Ismail, Mazlina and Joshi, Kroopa and Maxwell, Gavin and Noursadeghi, Mahdad and Riddell, Natalie and Ruehl, Tabea and Turner, Carolin T. and Uddin, Imran and Chain, Benny},
year = 2017,
month = oct,
journal = {Frontiers in Immunology},
volume = {8},
pages = {1267},
issn = {1664-3224},
doi = {10.3389/fimmu.2017.01267},
urldate = {2026-03-16}
}
@article{gierlinskiStatisticalModelsRNAseq2015,
title = {Statistical Models for {{RNA-seq}} Data Derived from a Two-Condition 48-Replicate Experiment},
author = {Gierli{\'n}ski, Marek and Cole, Christian and Schofield, Piet{`a} and Schurch, Nicholas J. and Sherstnev, Alexander and Singh, Vijender and Wrobel, Nicola and Gharbi, Karim and Simpson, Gordon and {Owen-Hughes}, Tom and Blaxter, Mark and Barton, Geoffrey J.},
year = 2015,
month = nov,
journal = {Bioinformatics},
volume = {31},
number = {22},
pages = {3625--3630},
issn = {1367-4811, 1367-4803},
doi = {10.1093/bioinformatics/btv425},
urldate = {2026-03-16},
copyright = {http://creativecommons.org/licenses/by/4.0/},
langid = {english},
file = {C:\Users\mcowley\Zotero\storage\HZNCNHX2\Gierliński et al. - 2015 - Statistical models for RNA-seq data derived from a two-condition 48-replicate experiment.pdf}
}
