Presentation is loading. Please wait.

Presentation is loading. Please wait.

Pairing T-cell Receptor Sequences using Pooling and Min-cost Flows

Similar presentations


Presentation on theme: "Pairing T-cell Receptor Sequences using Pooling and Min-cost Flows"— Presentation transcript:

1 Pairing T-cell Receptor Sequences using Pooling and Min-cost Flows
Tyler Daddio & Ion Mandoiu Computer Science & Engineering Department University of Connecticut

2 Class I endogenous antigen presentation

3 Interaction between TCR, peptide and MHC
Bernhard Knapp et al. Brief Bioinform 2015;16:

4 Somatic rearrangement of T-cell receptor genes
Potential TCR repertoire diversity: 1015

5 T-cell selection in thymus
Estimated TCR repertoire diversity after selection: ~2x107

6 T-cell activation and proliferation

7 T-cell activation and proliferation

8 T-cell activation and proliferation

9 Personalized cancer immunotherapy
Ton N. Schumacher, and Robert D. Schreiber Science 2015;348:69-74

10 Cancer vaccine design Which Epitopes go into the vaccine?
Epitope expression Expression level/pMHC abundance Cancer clone coverage MHC binding Binding affinity vs. differential agretopic index Diversity across the individual’s MHC alleles TCR binding TCR-pMHC binding affinity/avidity TCR precursor frequency

11 Single chain TCR sequencing
Klinger et al. 2013

12 Paired TCRs by single cell RNA-Seq assembly
Stubbington et al. Nature Methods 13, 329–332 (2016)

13 Paired TCRs by multiplexed emulsion PCR
European Journal of Immunology Volume 43, Issue 9, pages , 26 JUN 2013 DOI: /eji

14 pairSeq protocol … … TCRA TCRB
Howie et al. Science Transl Med 7(301), pp.301ra131 (2015)

15 Bipartite graph model Edge weights = hamming distance b/w well sets
Node capacities = #sequences with that well set Dummy nodes (empty well sets) to for undetected sequences

16 Perfect bipartite b-matching problem reduced to min-cost max-flow
Cost-scaling algo from LEMON library (Desz, Juttner, and Kovacz, 2011)

17 Graph sparsification Efficient implementation using multi-index hashing (Norouzi, Punjani, and Fleet, arXiv: )

18 Datasets Simulated datasets starting from experimental data from (Howie et al.)

19 Effectiveness of graph sparsification

20 Phase transition in b-matching cost

21 Pairing precision

22 Pairing recall

23 Pairing F-measure

24 Ongoing Work Experiments on additional datasets
Algorithm parallelization for multi-core/GPUs Integration with pMHC-TCR docking algorithms

25 Acknowledgements Holster First Year Project scholarship
Castleman term professorship in Engineering Innovation NSF award DBI


Download ppt "Pairing T-cell Receptor Sequences using Pooling and Min-cost Flows"

Similar presentations


Ads by Google