1 of 63

Analysis of Single-Cell T-Cell Receptor Sequencing

Kelly Street

Assistant Professor

Division of Biostatistics

Jan 12, 2023

2 of 63

What is a T cell?

3 of 63

T Cell Structure

4 of 63

Why should we be interested?

5 of 63

Motivation

6 of 63

Motivation

7 of 63

Motivation

8 of 63

Motivation

9 of 63

Motivation

10 of 63

Application: Renal Cell Carcinoma

Figure 1. Single-cell profiling of clear cell renal cell carcinoma

11 of 63

T Cell Exhaustion

Figure 3. CD8+ T cell trajectory analysis reveals increased terminal exhaustion with advancing disease stage

12 of 63

Trajectory Analysis

Figure 3. CD8+ T cell trajectory analysis reveals increased terminal exhaustion with advancing disease stage

13 of 63

TCR Sequencing

Figure 4. TCR analysis reveals lower diversity in terminally exhausted T cells

14 of 63

T Cell Structure

15 of 63

cellranger vdj

filtered_contig_annotations.csv

TCR Sequencing

16 of 63

barcode is_cell contig_id high_confidence length chain v_gene

TCGGGACAGACTAAGT-4 True TCGGGACAGACTAAGT-1_c.. True 826 TRB TRBV5-6

TCGGGACAGACTAAGT-4 True TCGGGACAGACTAAGT-1_c.. True 697 TRA TRAV6

TCGGGACAGACTAAGT-4 True TCGGGACAGACTAAGT-1_c.. True 695 TRA TRAV1-2

CTGATAGTCTTGTACT-4 True CTGATAGTCTTGTACT-1_c.. True 707 TRB TRBV6-6

CTGATAGTCTTGTACT-4 True CTGATAGTCTTGTACT-1_c.. True 851 TRA TRAV6

d_gene j_gene c_gene full_length productive cdr3 cdr3_nt reads

None TRBJ1-1 TRBC1 True True CASSPGDTEAFF TGTGCCAGCAGCCCGGGCGA.. 58796

None TRAJ34 TRAC True True CALDGPGNTDKLIF TGTGCTCTAGACGGCCCAGG.. 44152

None TRAJ33 TRAC True None None None 32494

TRBD2 TRBJ2-7 TRBC2 True True CASSYSTVYEQYF TGTGCCAGCAGTTACTCTAC.. 54812

None TRAJ39 None False None None None 478

umis raw_clonotype_id raw_consensus_id type sample

12 clonotype9 clonotype9_consensus_1 TCR S11_N

5 clonotype9 clonotype9_consensus_2 TCR S11_N

5 clonotype9 None TCR S11_N

9 clonotype1 clonotype1_consensus_2 TCR S11_N

17 clonotype1 None TCR S11_N

17 of 63

Cells can be ambiguous

chain productive cdr3

TRB True CASDIGGESYNELTF

TRA True CIPRFAFGPLTVF

18 of 63

Cells can be ambiguous

chain productive cdr3

TRB True CASDIGGESYNELTF

TRA True CIPRFAFGPLTVF

✔

19 of 63

Cells can be ambiguous

chain productive cdr3

TRB True CASDIGGESYNELTF

TRA True CIPRFAFGPLTVF

✔

chain productive cdr3

TRB True CSADISGSSYNEQFF

TRA True CIVRVAFGQNFVF

TRA None CAPSFSGNTPLVF

✔

20 of 63

Cells can be ambiguous

chain productive cdr3

TRB True CASDIGGESYNELTF

TRA True CIPRFAFGPLTVF

✔

chain productive cdr3

TRB True CSADISGSSYNEQFF

TRA True CIVRVAFGQNFVF

TRA None CAPSFSGNTPLVF

✔

chain productive cdr3

TRB True CSADISGSSYNEQFF

TRA True CIVRVAFGQNFVF

TRA True CAPSFSGNTPLVF

?

21 of 63

Cells can be ambiguous

chain productive cdr3

TRB True CASDIGGESYNELTF

TRA True CIPRFAFGPLTVF

✔

chain productive cdr3

TRB True CSADISGSSYNEQFF

TRA True CIVRVAFGQNFVF

TRA None CAPSFSGNTPLVF

✔

chain productive cdr3

TRB True CSADISGSSYNEQFF

TRA True CIVRVAFGQNFVF

TRA True CAPSFSGNTPLVF

?

chain productive cdr3

TRB True CASSQEAGTSGNTIYF

TRB None CSAMVRPSGNTNKLTF

?

22 of 63

Cells can be ambiguous

chain productive cdr3

TRB True CASDIGGESYNELTF

TRA True CIPRFAFGPLTVF

✔

chain productive cdr3

TRB True CSADISGSSYNEQFF

TRA True CIVRVAFGQNFVF

TRA None CAPSFSGNTPLVF

✔

chain productive cdr3

TRB True CSADISGSSYNEQFF

TRA True CIVRVAFGQNFVF

TRA True CAPSFSGNTPLVF

?

chain productive cdr3

TRB True CASSQEAGTSGNTIYF

TRB None CSAMVRPSGNTNKLTF

?

chain productive cdr3

TRB True CASSWGLGTEAFF

IGL None None

TRA None None

TRA True CALSGRGEGGSEKLVF

TRA True CAGLDTGTASKLTF

IGH None None

Multi None None

IGK None None

?

23 of 63

# alpha chains

# beta chains

Cells can be ambiguous

24 of 63

Motivation

25 of 63

Cells can be ambiguous

chain productive cdr3

TRB True CASDIGGESYNELTF

TRA True CIPRFAFGPLTVF

chain productive cdr3

TRB True CSADISGSSYNEQFF

TRA True CIVRVAFGQNFVF

TRA None CAPSFSGNTPLVF

chain productive cdr3

TRB True CSADISGSSYNEQFF

TRA True CIVRVAFGQNFVF

TRA True CAPSFSGNTPLVF

chain productive cdr3

TRB True CASSQEAGTSGNTIYF

TRB None CSAMVRPSGNTNKLTF

chain productive cdr3

TRB True CASSWGLGTEAFF

IGL None None

TRA None None

TRA True CALSGRGEGGSEKLVF

TRA True CAGLDTGTASKLTF

IGH None None

Multi None None

IGK None None

26 of 63

Cells can be ambiguous

chain productive cdr3

TRB True CASDIGGESYNELTF

TRA True CIPRFAFGPLTVF

chain productive cdr3

TRB True CSADISGSSYNEQFF

TRA True CIVRVAFGQNFVF

TRA None CAPSFSGNTPLVF

chain productive cdr3

TRB True CSADISGSSYNEQFF

TRA True CIVRVAFGQNFVF

TRA True CAPSFSGNTPLVF

chain productive cdr3

TRB True CASSQEAGTSGNTIYF

TRB None CSAMVRPSGNTNKLTF

chain productive cdr3

TRB True CASSWGLGTEAFF

IGL None None

TRA None None

TRA True CALSGRGEGGSEKLVF

TRA True CAGLDTGTASKLTF

IGH None None

Multi None None

IGK None None

27 of 63

Clonotype Identification

Ambiguous clonotype

Unique clonotype

beta

alpha

Many cells have a unique, identifiable clonotype

Rare clonotypes are clinically relevant

28 of 63

Do these cells share a clonotype?

Clonotype Identification

beta

alpha

common beta

common alpha

29 of 63

Do these cells share a clonotype?

Clonotype Identification

beta

alpha

common beta

common alpha

30 of 63

Do these cells share a clonotype?

Clonotype Identification

beta

alpha

common beta

common alpha

31 of 63

Do these cells share a clonotype?

Clonotype Identification

beta

alpha

common beta

common alpha

32 of 63

TCR analysis

Clonotype information was derived from the Cell Ranger V(D)J Annotation pipeline v2.1.0. All assembled contigs were filtered to retain only those that were assigned a raw clonotype ID and categorized as being both full length and productive. Cell barcodes were then filtered to retain only those with two contigs, categorized as one TCR alpha chain and one TCR beta chain. Each clonotype was assigned a unique identifier, consisting of the predicted amino acid sequences of the CDR3 regions of these two chains, which was used to match clonotypes across samples.

Option 1: Remove ambiguity

33 of 63

Option 2: Everything’s a clonotype

beta

alpha

common beta

common alpha

Clonotype 1:

Clonotype 2:

34 of 63

Option 2: Everything’s a clonotype

beta

alpha

common beta

common alpha

Clonotype 1:

Clonotype 2:

35 of 63

Option 3: Big Tent clonotypes

36 of 63

Clonotype Identification

Clonotype 1:

Clonotype 2:

Ambiguous:

beta

alpha

common beta

common alpha

37 of 63

Proportional Assignment (E-M)

66.6% Clonotype 1

33.3% Clonotype 2

Clonotype 1:

Clonotype 2:

beta

alpha

common beta

common alpha

38 of 63

Axiom: a clonotype is defined by a single alpha+beta combination
This is a classic Expectation-Maximization problem (see: RSEM)

multi-mapped reads : gene transcripts :: ambiguous T cells : clonotypes

Proportional Assignment (E-M)

read

transcripts

T cell

clonotypes

50 - 50

39 of 63

Axiom: a clonotype is defined by a single alpha+beta combination
This is a classic Expectation-Maximization problem (see: RSEM)

multi-mapped reads : gene transcripts :: ambiguous T cells : clonotypes

80 - 20

Proportional Assignment (E-M)

read

transcripts

T cell

clonotypes

40 of 63

Axiom: a clonotype is defined by a single alpha+beta combination
This is a classic Expectation-Maximization problem (see: RSEM)

multi-mapped reads : gene transcripts :: ambiguous T cells : clonotypes

Proportional Assignment (E-M)

read

transcripts

T cell

clonotypes

100 - 00

41 of 63

Same Data, More Cells

42 of 63

Same Data, More Cells

Unique Counts

log1p Unique Counts

E-M Counts

log E-M Counts

43 of 63

Total clonotypes
(normalized) Shannon entropy
(inverse) Simpson Index

(currently) limited:

Chao1
Chao-Bunge
breakaway
breakaway_nof1

Quantifying (alpha) Diversity

44 of 63

Improved Signal

Figure 4. TCR analysis reveals lower diversity in terminally exhausted T cells

45 of 63

Improved Signal

Unique

E-M

46 of 63

Simulation

47 of 63

Bioconductor Package

48 of 63

Mercedeh Movassagh

Jill Lundell

David Braun

Mingzhi Ye

The Team

49 of 63

50 of 63

Bonus Content:

Quantifying Plasticity

in Cancer

Kelly Street

Assistant Professor

Division of Biostatistics

Jan 12, 2023

51 of 63

What about DNA methylation?

Plasticity

Shen S, Clairambault J. Cell plasticity in cancer cell populations. F1000Res. 2020 Jun 22;9:F1000 Faculty Rev-635. doi: 10.12688/f1000research.24803.1. PMID: 32595946; PMCID: PMC7309415.

52 of 63

Tumor samples from colorectal cancer patients
Paired samples from left and right sides of tumors
Gene-level variability is the average difference in % methylation between in the gene body

Methylation Variability

53 of 63

Expression Variability

54 of 63

Expression Variability

55 of 63

Expression Variability

56 of 63

Plasticity

GSE97693

57 of 63

Plasticity

GSE81861

58 of 63

Plasticity

GSE201348

59 of 63

David Braun

Plasticity

HCRN

60 of 63

Plasticity

GSE97693

61 of 63

DepMap “essential” genes

Plasticity

GSE97693

62 of 63

Kimberly Siegmund

Darryl Shibata

The Team

1 of 63

2 of 63

3 of 63

4 of 63

5 of 63

6 of 63

7 of 63

8 of 63

9 of 63

10 of 63

11 of 63

12 of 63

13 of 63

14 of 63

15 of 63

16 of 63

17 of 63

18 of 63

19 of 63

20 of 63

21 of 63

22 of 63

23 of 63

24 of 63

25 of 63

26 of 63

27 of 63

28 of 63

29 of 63

30 of 63

31 of 63

32 of 63

33 of 63

34 of 63

35 of 63

36 of 63

37 of 63

38 of 63

39 of 63

40 of 63

41 of 63

42 of 63

43 of 63

44 of 63

45 of 63

46 of 63

47 of 63

48 of 63

49 of 63

50 of 63

51 of 63

52 of 63

53 of 63

54 of 63

55 of 63

56 of 63

57 of 63

58 of 63

59 of 63

60 of 63

61 of 63

62 of 63

63 of 63