Genome-Wide SNP Discovery from de novo Assemblies

Genome
-
Wide SNP Discovery from
de novo

Assemblies

of

Pepper (
Capsicum annuum
) Transcriptomes

Hamid Ashrafi
1
, Jiqiang Yao
2
, Kevin Stoffel
1
, Sebastian R. Chin
-
Wo
3
, Theresa Hill
1
, Alexander Kozik
3

and Allen Van Deynze
1


1

Department of Plant Sciences, Seed Biotechnology Center, University of California, Davis, CA 95616


2

Interdisciplinary Center for Biotechnology Research (ICBR), University of Florida, Gainesville, FL 32610


3
Genome Center, University of California, Davis, CA 95616

Background and Significance


To obtain

as many transcribed genes
as possible,
peppers were
sampled from different
,

cultivars, tissues
at multiple stages of growth
and development.


To discover putative SNPs among three sampled pepper cultivars by
sequencing transcriptomes using
Illumina

Genome Analyzer.


To annotate the transcriptome sequence in order to have an insight into
pepper biological processes.


To use annotated gene
s for QTL analysis and
candidate gene
discovery.

Objectives

Materials and Methods

Results

Conclusions

References

Acknowledgments

Plant

Materials

and

cDNA

Library

Preparation


The

seed

of

three

pepper

(
C
.

annuum)

lines

‘CM
334
,’

‘
Maor
’

and

‘Early

Jalapeño’

were

planted
.


Three

cDNA

libraries

(one

from

each

pepper

variety)

were

prepared

using

pooled

RNA

that

was

extracted

from

4

tissues
:

root,

young

leaf,

flower

and

fruit

using

Qiagen

RNeasy

Mini

Kit

(Qiagen

Valencia

CA,

USA)
.

Fruit

tissues

were

collected

in

different

developmental

stages
;

5
,

10
,

and

20

days

post

pollination

developing

fruit,

breaker

and

ripe

fruit
.



The

libraries

were

constructed

by

shearing

cDNAs

and

300
‐
350

bp

fragments

were

selected

on

gels
.


The

libraries

were

normalized

using

a

double
-
stranded

nuclease

protocol
.


The

cDNA

libraries

were

sequenced

using

Illumina

Genome

Analyzer

IIx

(
GAIIx
)

(Illumina

Inc
.
,

San

Diego,

CA)

for

80
-
120

cycles

at

UC

Davis

Genome

Center

core

facility
.


De

Novo

Assembly

of

NGS

Sequences


The

NGS

data

(
GAIIx
)

went

through

our

standard

preprocessing

pipeline,

developed

at

UC

Davis

(Kozik,

A,

2010
)
.


Velvet

(
Zerbino

and

Birney,

2008
)

and

CLC

(CLCBIO,

2010
)

software

packages

were

used

to

assemble

the

sequences
.


CAP
3

was

used

to

make

the

final

assembly

of

three

assemblies
.














Trimmed reads

Min 40nt
–

Max 85nt

Trimmed reads

Min 25nt
–

Max 60nt

One iteration of CLC

assembly with all reads

One iteration of CLC

assembly with all reads

One iteration of CLC

assembly with all reads

Velvet Assembler

Early Jalape
ñ
o

31

35

41

31

35

41

Trimmed reads

Min 40nt
–

Max 85nt

Trimmed reads

Min 25nt
–

Max 60nt

All K
-
mer

assemblies,

assembled with CAP3

Ma
or

31

35

41

Trimmed reads

Min 40nt
–

Max 85nt

Trimmed reads

Min 25nt
–

Max 60nt

31

35

41

All K
-
mer

assemblies,

assembled with CAP3

CM334

31

35

41

31

35

41

Trimmed reads

Min 40nt
–

Max 85nt

Trimmed reads

Min 25nt
–

Max 60nt

All K
-
mer

assemblies,

assembled with CAP3

Velvet K
-
mers

CLC Assembler

Trimmed reads

Min 40nt
–

Max 85nt

Trimmed reads

Min 25nt
–

Max 60nt

Trimmed reads

Min 40nt
–

Max 85nt

Trimmed reads

Min 25nt
–

Max 60nt

Velvet Assembler

CLC Assembler

Velvet Assembler

CLC Assembler

CM334
assembly made with CAP3

Early Jalape
ñ
o

assembly made with CAP3

Ma
or

assembly
made with CAP3

+

+

+

Pepper final assembly made with CAP3

(Reference Sequence)

Assembly

Statistics



No
.

of

Contigs


Total

nt


N
50

CM
334



83
,
113


84
,
792
,
180

1
,
488

Early

Jalapeño



82
,
614


84
,
973
,
865

1
,
488

Maor


76
,
375


79
,
383
,
673

1
,
526


Pepper

assembly


123
,
261


135
,
019
,
787

1
,
647

(CM
334
,EJ

and

Maor
)



















Annotation


A

total

of

63
,
202

contigs

(
51
.
3
%
)

had

at

least

one

hit

in

the

non
-
redundant

database

of

GenBank

with

an

average

length

of

1
,
495

nucleotides
.


Contigs

with

a

hit,

covered

94
.
5

M

bases

(
70
%
)

of

the

total

assembly
.


A

total

of

60
,
055

(
48
.
7
%
)

contigs

that

did

not

have

any

hit

in

the

GenBank

were

on

average

674

nucleotide

long

and

covering

40
.
5

M

bases

(
30
%
)

of

the

total

assembly
.


Based

on

all

results

of

BLASTX,

Vitis

vinifera
,

Arabidopsis

thaliana

and

Oryza

sativa

were

the

top

three

species

in

the

blast

hits


(Fig

3
)
.


Mapping

step

of

Blast
2
GO

resulted

in

identification

of

37
,
918

(
30
.
7
%
)

contigs

with

Gene

Ontology

(GO)

terms
.


Biological

Processes

(BP)

at

different

GO

levels

were

generated
.

Fig

4

shows

the

BP

at

level

2
.

For

each

BP

number

of

annotated

sequences

are

shown

in

Fig

5
.



Kegg

maps

for

150

biological

pathways

were

generated

and

contigs

within

each

pathways

were

determined
.

For

instance,

Fig

6

depicts

Kegg

map

of

Pyrimidine

Metabolism

pathway
.


SNP

discovery



A

total

of

22
,
863

putative

SNPs

within

11
,
869

contigs

were

identified

by

our

SNP

discovery

pipeline
.


The

contigs

with

identified

putative

SNPs

comprised

23
,
794

kb

(
17
.
6
%
)

of

pepper

transcriptomes

assembly
.


On

average

1

SNP

per

1040

bp

of

exonic

regions

of

pepper

genome

was

identified
.


Assembly

of

transcriptomes

of

three

pepper

cultivars,

increased

the

total

assembled

bases

by

50
%
.



The

present

pepper

transcriptome

assembly

represents

~
4
%

of

pepper

genome

(
3500

Mb)
.


We

demonstrated

that

for

the

plants

for

which

the

genome

sequences

are

not

available

yet,

the

transcriptome

assembly

is

an

alternate

approach

SNP

calling
.



Annotation

of

51
%

of

contigs

or

70
%

of

total

assembled

bases

indicates

that

~
49
%

of

contigs

are

small

contigs

that

are

covering

the

remaining

30
%

of

unannotated

sequences
.







Conesa,

A
.
,

S
.

Götz
,

et

al
.

(
2005
)
.

"Blast
2
GO
:

a

universal

tool

for

annotation,

visualization

and

analysis

in

functional

genomics

research
.
"

Bioinformatics

21
(
18
)
:

3674
-
3676
.

Kozik
,

A

(
2010
)
.
Tool

to

process

and

manipulate

Illumina

sequences)
.

http
:
//code
.
google
.
com/p/atgc
-
illumina/downloads/list
”)
.


Li,

H
.

and

R
.

Durbin

(
2009
)
.

"Fast

and

accurate

short

read

alignment

with

Burrows
–
Wheeler

transform
.
"

Bioinformatics

25
(
14
)
:

1754
-
1760
.

Li,

H
.
,

B
.

Handsaker
,

et

al
.

(
2009
)
.

"The

Sequence

Alignment/Map

format

and

SAMtools
.
"

Bioinformatics

25
(
16
)
:

2078
-
2079
.

Zerbino
,

D
.

and

E
.

Birney

(
2008
)
.

"Velvet
:

algorithms

for

de

novo

short

read

assembly

using

de

Bruijn

graphs
.
"

Genome

Res

18
:

821

-

829
.





Molecular breeding of pepper (
Capsicum spp.
) has been hampered by the paucity of
molecular markers.


This is primarily due to lack of availability of the pepper genome sequence and
limited available sequence resources.


In recent years with the more cost effective sequencing technologies such as
Illumina
, sequencing of expressed genes (transcriptomes), gene discovery and
allele mining is no longer insurmountable.


In order to exploit the speed and scale of data from new sequencing technologies
and in an effort to enrich the sequence resources of pepper, we sequenced
transcriptome sequences (RNA
-
seq
) of three pepper lines: Maor, Early Jalapeño
(EJ) and
Criollo

de
Morelos 334 (CM334).


We selected a wide range of tissues to represent as many expressed genes as
possible.


The reference sequence was constructed from >200 million
Illumina

reads

(80
-
120
nt
) using a combination of Velvet, CLC and CAP3 software packages.


BWA (Li and Durbin, 2009),
SAMtools

(Li et al, 2009b) and in
-
house Perl scripts
were used to identify SNPs among three pepper lines.


The SNPs were filtered to be 100
bp

apart from any putative

intron
-
exon junctions
as well as adjacent SNPs. After filtering >22,000 high quality putative SNPs were
identified and
bioinformatically

mapped to pepper genetic maps.


The reference sequence was annotated by Blast2Go software (Conesa et al, 2005).

The

authors

would

like

to

thank

Enza

Zaden
,

Nunhems,

Rijk

Zwaan,

Syngenta,

Vilmorin

and

UC

Discovery

program

for

the

financial

support
.

We

also

would

like

to

thank

sequencing

facility

of

UC

Davis

Genome

Center

and

Bioinformatics

core

facility

to

provide

us

the

servers

and

computational

power
.

The

annotation

would

not

be

possible

without

collaboration

with

Dr
.

R

Michelmore’s

laboratory
.





SNP

Discovery

Pipeline


BWA

was

used

to

map

all

the

reads

of

three

genotypes

individually

to

the

Pepper

final

transcriptome

assembly
.


SAMtools

was

use

to

make

the

pileups

of

each

cultivar

and

discover

the

difference

within

each

cultivar

with

reference

sequence
.


Indels

were

screened

out

of

pileup

files
.



Intron
-
exon

junction

positions

were

inferred

in

the

reference

sequence

based

on

Arabidopsis

gene

models

using

intron

finder

of

Solanaceae

Genome

Network

website

(SGN)
.


In
-
house

Perl

scripts

were

used

to

create

allele

call

table

of

all

three

genotypes,

the

SNPs

were

filtered

against

adjacent

SNPs

and

identified

Intronic

regions
.


Sequences

surrounding

the

SNPs

(
100

base

on

each

side)

were

extracted

from

the

reference

sequence

to

design

assays
.

Annotation

of

Reference

Sequence


Blast2Go program was used to annotate the reference sequence, obtain the statistics and generate Kegg maps(http://www.genome
.jp
/kegg/pathway.html).





Fig
.

3

Fig
.

4

Fig
.

5

Fig
.

2

Distribution

of

contig

length

in

pepper

transcriptome

assembly


N50=1647

Mean=1095


Max=19,089

Min=265

Fig
.

6

Kegg

map

of

Pyrimidine

Metabolism

Fig
.

1

De

Novo

assembly

of

pepper

transcriptomes