1 of 34

Advanced immunoinformatics methods as supportive tools in antibody drug and vaccine design

Yana Safonova

UCSD

2 of 34

Adaptive immune receptor repertoires (AIRR)

2

Antibody

T-cell receptor

Cancer immunotherapy

HIV vaccine

Infection disease treatment

Autoimmune studies

3 of 34

Rep-seq technologies

3

V

D

J

Left read

Right read

~ 400 nt

4 of 34

Rep-seq technologies

4

V

D

J

Left read

Right read

~ 400 nt

Read covering VDJ

5 of 34

Rep-seq technologies

5

V

D

J

Left read

Right read

~ 400 nt

Read covering VDJ

6 of 34

Computational challenges

6

Selected B cells

Immunoglobulin RNAs

7 of 34

Computational challenges

7

Selected B cells

Error-prone Rep-seq reads

8 of 34

Computational challenges

8

Selected B cells

Antibody repertoire

IgReC

tool for construction of AIRR from Rep-seq data

​

yana-safonova.github.io/ig_repertoire_constructor

Shlemov et al, Journal of Immunology, 2017 (December)

comparison of repertoire construction methods

​

9 of 34

Immunoinformatics pipeline

9

Recombination and SHM statistics

Analysis of immune response dynamics

Clonal analysis

Population of Ig loci

10 of 34

Immunoinformatics pipeline

10

Recombination and SHM statistics

Analysis of immune response dynamics

Clonal analysis

Population of Ig loci

11 of 34

VDJ classification

  • Finds the closest V, D, J from database of germline segments
  • Provides basic understanding of AIRR diversity

​

​

​

​

Quality of VDJ classification directly depends on quality of germline segments

11

V1

V2

V3

V4

J1

J2

J3

J4

D1

D2

D3

D4

J1

J2

J3

J4

V1

V2

V3

V4

12 of 34

Finding novel IGHV segments

12

✤ ✺

✤ ✺

✤ ✺

Closest known V segment

differences from the closest segment from the database

13 of 34

Finding novel IGHV segments

13

Novel V segments:

✤ ✺

✤ ✺

✤ ✺

✤ ✺

✤ ✺

✪ ❃ ❇

✪ ❃ ❇

✤ ✺

Closest known V segment

14 of 34

Finding novel IGHV segments

14

✤ ✺

✤ ✺

✤ ✺

✤ ✺

✤ ✺

✪ ❃ ❇

✪ ❃ ❇

✪ ❃ ❇

✪ ❃ ❇

✪ ❃ ❇

✤ ✺

✪ ❃ ❇

Closest known V segment

Novel V segments:

15 of 34

Finding novel IGHV segments

15

✤ ✺

✤ ✺

✤ ✺

✤ ✺

✤ ✺

✪ ❃ ❇

✪ ❃ ❇

✪ ❃ ❇

✪ ❃ ❇

✪ ❃ ❇

Our analysis revealed:

  • > 100 segments with < 4 mismatches
  • 19 segments with 4-6 mismatches
  • 17 segments with by 7-13 mismatches
  • 6 segments with 14-89 mismatches

​

Closest known V segment

✤ ✺

✪ ❃ ❇

Novel V segments:

16 of 34

Finding IGHD segments

16

GAGCGAGCGGA

CCCCGAGCGCATA

ATTGCGAGCGCCCC

GCGAGCGCAGA

GCGAGCGCA

V1

GAGCGAGCGGA

J1

V2

CCCCGAGCGCATA

J2

V3

ATTGCGAGCGCCCC

J3

V4

GCGAGCGCAGA

J4

Naive CDR3s

Novel D segments

17 of 34

Finding IGHD segments

17

GAGCGAGCGGA

CCCCGAGCGCATA

ATTGCGAGCGCCCC

GCGAGCGCAGA

GCGAGCGCA

Naive CDR3s

Novel D segments

We detected:

  • 10 novel D segments
  • 100 mosaic D segments!

V1

GAGCGAGCGGA

J1

V2

CCCCGAGCGCATA

J2

V3

ATTGCGAGCGCCCC

J3

V4

GCGAGCGCAGA

J4

18 of 34

Unequal crossover contributes to diversity of Ig loci

18

V2

V3

V1

V2

V3

V1

father chr.

mother chr.

19 of 34

Unequal crossover contributes to diversity of Ig loci

19

V2

V3

V1

V2

V3

V1

father chr.

mother chr.

20 of 34

Unequal crossover contributes to diversity of Ig loci

20

V2

V3

V1

V2

V3

V1

V21

V2

V3

V12

V3

V1

father chr.

mother chr.

21 of 34

Antibodies gain mutations during maturation

21

V

D

J

22 of 34

Antibodies gain mutations during maturation

​

22

✪

✺

✿

✷

✳

✽

✾

✾

23 of 34

Antibodies gain mutations during maturation

​

23

V

✪

✺

✿

✷

✳

✽

✾

✾

VDJ classification is applied for computation of SHMs in antibodies

24 of 34

Antibodies gain mutations during maturation

​

24

V

✪

✺

✿

✷

✳

✽

✾

✾

❄

❈

✷

All variations of germline segments would be considered as SHMs

25 of 34

Antibodies gain mutations during maturation

​

25

✪

✺

✿

✷

✳

✽

✾

✾

❄

❈

✷

All variations of germline segments would be considered as SHMs

V

❄

❈

✷

26 of 34

Clonal tree

26

Clonal trees represent evolutionary development of antibody repertoire, including:

  • somatic hypermutagenesis
  • selection
  • clonal expansion

27 of 34

Clonal trees for various repertoires

27

healthy individual

flu-vaccinated individual

HIV-infected individual

28 of 34

Clonal tree as a history of SHMs

28

V

J

✵

✵

❂

❂

❂

✽

✽

❈

✹

✹

❈

❈

❈

❂

A

B

C

D

✵

❈

❂

E

✦

✦

Naive SHM counting:

3 × ✵

5 × ❂

2 × ✽

2 × ✹

2 × ✦

5 × ❈

29 of 34

Clonal tree as a history of SHMs

29

V

J

✵

✵

❂

❂

❂

✽

✽

❈

✹

✹

❈

❈

❈

❂

A

B

C

D

❂

❈

✵

❈

❂

E

✦

✦

SHMs in clonal tree:

A

D

B

C

✵

✹

✽

✹ + ✦

E

✦

Naive SHM counting:

3 × ✵

5 × ❂

2 × ✽

2 × ✹

2 × ✦

5 × ❈

1 × ✵

1 × ✽

2 × ✹

2 × ✦

30 of 34

Clonal tree as a history of SHMs

30

V

J

✵

✵

❂

❂

❂

✽

✽

❈

✹

✹

❈

❈

❈

❂

A

B

C

D

✵

❈

❂

E

✦

✦

❈

❂

✦

✹

Drug candidate:

A

D

B

C

✵

✹

✽

✹ + ✦

E

✦

❂

❈

SHMs in clonal tree:

1 × ✵

1 × ✽

2 × ✹

2 × ✦

31 of 34

Ideal pipeline

31

❈

❂

✦

✹

Drug candidate

32 of 34

Ideal pipeline

32

✤ ✺

✪ ❃ ❇

✦ ✩

✮❆✿

Novel segments

Clonal trees

❈

❂

✦

✹

Drug candidate

33 of 34

Ideal pipeline

33

Naive Rep-seq

Mutated Rep-seq

✤ ✺

✪ ❃ ❇

✦ ✩

✮❆✿

Novel segments

Clonal trees

❈

❂

✦

✹

Drug candidate

34 of 34

Acknowledges

34

isafonova@eng.ucsd.edu

Pavel Pevzner

Timofey Prodanov

Andrey Slabodkin

Andrey Bzikadze

Alex Shlemov

Sergey Bankevich