NGS cancer genomics data processing and analysis

NGS cancer genomics data processing and
analysis

Somak Roy, MD


Clinical fellow

Division of Urologic Surgical Pathology

University of Pittsburgh Medical Center

Outline

http://www.genome.gov/sequencingcosts/

Background

Epigenetic
profiling

NGS

Gene
fusion
detection

Mutation
profiling

Copy
number
variations

Structural
variants

Application in Cancer Genomics

NGS

in clinical domain

Sequence the sample DNA to obtain a string of characters
(ATGC)

Compare the obtained sequence to the reference
sequence (expected normal)

Any deviation from the reference (single or multiple
base(s)) is a variant.

Theme of DNA Sequencing

Evolution of Sequencing

Robison. Nat
Biotechnol

2011;29:805
-
7

Rothberg et al. Nature.
2011;475:348
-
52

Semiconductor Sequencing

Arch
Pathol

Lab Med. 2012;136:000
–
000;
doi
: 10.5858/arpa.2012
-
0107
-
RA

Optics
-
based Sequencing

NGS data processing elements

Signal processing

Raw signal

Normalization

Base calling

Quality control

FASTQ / unaligned BAM

Signal Processing
–

Non
-
optical

CCGCTAGCTATATTATATCGGGGCCCTAGATAGCTAGATATAGAGGGCTCTAGAGATCGATAGCTAGAG

CTAGCTCG
CCGGGGCCCTAGAGTATATTATAGGCTCTAGAGATCGATAGCTGATAGCTAGATATAAGAG

ATATAAGCGCGGCTCGATCGGTCTAGAGAGGCCCTAGAGTATATTA
CTAGCT
TAAGCTGATAGCTAGAG

CCGCTAGCTATATTATATCGGGGCCCTAGATAGCTAGATATAGAGGGCTCTAGAGATCGATAGCTAGAG

CTAGCTCG
CCGGGGCCCTAGAGTATATTATAGGCTCTAGAGATCGATAGCTGATAGCTAGATATAAGAG

ATATAAGCGCGGCTCGATCGGTCTAGAGAGGCCCTAGAGTATATTA
CTAGCT
TAAGCTGATAGCTAGAG

CCGCTAGCTATATTATATCGGGGCCCTAGATAGCTAGATATAGAGGGCTCTAGAGATCGATAGCTAGAG

CTAGCTCG
CCGGGGCCCTAGAGTATATTATAGGCTCTAGAGATCGATAGCTGATAGCTAGATATAAGAG

ATATAAGCGCGGCTCGATCGGTCTAGAGAGGCCCTAGAGTATATTA
CTAGCT
TAAGCTGATAGCTAGAG

CCGCTAGCTATATTATATCGGGGCCCTAGATAGCTAGATATAGAGGGCTCTAGAGATCGATAGCTAGAG

CTAGCTCG
CCGGGGCCCTAGAGTATATTATAGGCTCTAGAGATCGATAGCTGATAGCTAGATATAAGAG

ATATAAGCGCGGCTCGATCGGTCTAGAGAGGCCCTAGAGTATATTA
CTAGCT
TAAGCTGATAGCTAGAG

Signal Processing
-

Optical

CCGCTAGCTATATTATATCGGGGCCCTAGATAGCTAGATATAGAGGGCTCTAGAGATCGATAGCTAGAG

CTAGCTCG
CCGGGGCCCTAGAGTATATTATAGGCTCTAGAGATCGATAGCTGATAGCTAGATATAAGAG

ATATAAGCGCGGCTCGATCGGTCTAGAGAGGCCCTAGAGTATATTA
CTAGCT
TAAGCTGATAGCTAGAG

CCGCTAGCTATATTATATCGGGGCCCTAGATAGCTAGATATAGAGGGCTCTAGAGATCGATAGCTAGAG

CTAGCTCG
CCGGGGCCCTAGAGTATATTATAGGCTCTAGAGATCGATAGCTGATAGCTAGATATAAGAG

ATATAAGCGCGGCTCGATCGGTCTAGAGAGGCCCTAGAGTATATTA
CTAGCT
TAAGCTGATAGCTAGAG

CCGCTAGCTATATTATATCGGGGCCCTAGATAGCTAGATATAGAGGGCTCTAGAGATCGATAGCTAGAG

CTAGCTCG
CCGGGGCCCTAGAGTATATTATAGGCTCTAGAGATCGATAGCTGATAGCTAGATATAAGAG

ATATAAGCGCGGCTCGATCGGTCTAGAGAGGCCCTAGAGTATATTA
CTAGCT
TAAGCTGATAGCTAGAG

CCGCTAGCTATATTATATCGGGGCCCTAGATAGCTAGATATAGAGGGCTCTAGAGATCGATAGCTAGAG

CTAGCTCG
CCGGGGCCCTAGAGTATATTATAGGCTCTAGAGATCGATAGCTGATAGCTAGATATAAGAG

ATATAAGCGCGGCTCGATCGGTCTAGAGAGGCCCTAGAGTATATTA
CTAGCT
TAAGCTGATAGCTAGAG

Signal Processing
-

Homopolymer

Take a Peak into FASTQ !

Header: Sequence ID, additional info

Sequence

Optional header

Quality score

Phred

Score /
Phred
-
like score

Q

=
-
10*log
10
p

Per Base
Call score

Take a Peak into FASTQ !

Q

=
-
10*log
10
p

30
=
-
10*log
10
(10
-
3
)

20
=
-
10*log
10
(10
-
2
)

What are these characters ?

ASCII format

67
=
-
10*log
10
(p=?)

Read

Pile
-
up

G
G
G
Depth of
Coverage

5x

3x

Var

(G) frequency =3/5 (60%)

Variant
frequency

Mapping, Assembly & Variant Identification

Mapping / Alignment

Next
-
Generation DNA Sequencing Informatics. Ed. Brown SM. Cold Spring Harbor Laboratory Press. 2013

Mapping / Alignment

Pabinger

et al. Briefings in Bioinformatics. Jan 2013

Mapping / Alignment
-

QC

Next
-
Generation DNA Sequencing Informatics. Ed. Brown SM. Cold Spring Harbor Laboratory Press. 2013

Variant identification

Variant identification

Pabinger

et al. Briefings in Bioinformatics. Jan 2013

Variant identification
-

QC

Next
-
Generation DNA Sequencing Informatics. Ed. Brown SM. Cold Spring Harbor Laboratory Press. 2013

Annotation

1522648G>A

44512584G>C

8124526T>A

2544856_2544860
AATGC

..

55124785GA>CC

Public / custom databases

Nomenclature

Biological implication (Gene, transcript and protein level)

Genotype
-
phenotype correlations

Prognostic implication

Predictive implication

Report level

Definition / explanation

1

Missense
SNVs
, insertions and deletions, occurring in the coding, untranslated and splice site
regions, which are of
known

clinical significance.

2

Missense
SNV
, insertions and deletions, occurring in the coding, untranslated and splice site
regions, which are of
uncertain

clinical significance.

3

Variant(s) not classifiable into
levels
1, 2, 4, and 5 automatically by program. Pathologist is required
to review it and reclassify appropriately based on available evidence.

4

Synonymous SNV, SNV(s) with established evidence of benign biological and/or clinical outcome
and intronic variants (except splice
site and deep
intronic

with known significance
).

5

Sequence variants which have been proved to be platform specific sequencing errors based on
repeated experimental
data,
sanger

sequence confirmation
and thorough manual review of
read
pile
-
ups
(
eg
. Using
IGV

viewer).

Variant Classification

Visualization

Result Reporting, Management and Sharing

Targeted
sequencing

Whole
exome

sequencing

Whole genome
sequencing