Cost-effective genome-wide estimation of allele frequencies from pooled DNA in Atlantic salmon (Salmo salar L.)
Full text
METHODOLOGY ARTICLE Open Access
Cos -e ec i e genome-wide es ima ion o allele
equencies om pooled DNA in A lan ic salmon
(Salmo sala L.)
Mikhail Oze o
1
, An i Vasemägi
2,7
, Vida Wenne ik
3
, Ee o Niemelä
4
, Se gey P uso
5
, Ma hew Ken
6
and Juha-Pekka Vähä
2*
Abs ac
Backg ound: New sequencing echnologies ha e emendously inc eased he numbe o known molecula
ma ke s (single nucleo ide polymo phisms; SNPs) in a a ie y o species. Concu en ly, imp o emen s o geno yping
echnology ha e now made i possible o e icien ly geno ype la ge numbe s o genome-wide dis ibu ed SNPs
enabling genome wide associa ion s udies (GWAS). Howe e , geno yping signi ican numbe s o indi iduals wi h
la ge numbe o SNPs emains p ohibi i ely expensi e o many esea ch g oups. A possible solu ion o his
p oblem is o de e mine allele equencies om pooled DNA samples, such ‘allelo yping’has been p esen ed as a
cos -e ec i e al e na i e o indi idual geno yping and has become popula in human GWAS. In his a icle we ha e
es ed he e ec i eness o DNA pooling o ob ain accu a e allele equency es ima es o A lan ic salmon
(Salmo sala L.) popula ions using an Illumina SNP-chip.
Resul s: In o al, 56 A lan ic salmon DNA pools om 14 popula ions we e analyzed on an A lan ic salmon SNP-chip
con aining p obes o 5568 SNP ma ke s, 3928 o which we e bi-allelic. We de eloped an e icien quali y con ol
il e which enables exclusion o loci showing high e o a e and mino allele equency (MAF) close o ze o. A e
applying mul iple quali y con ol il e s we ob ained allele equency es ima es o 3631 bi-allelic loci. We obse ed
high conco dance ( > 0.99) be ween allele equency es ima es de i ed om indi idual geno yping and DNA
pools. Ou esul s also indica e ha e en ela i ely small DNA pools (35 indi iduals) can p o ide accu a e allele
equency es ima es o a gi en sample.
Conclusions: Despi e o highe le el o a ia ion associa ed wi h a ay eplica es compa ed o pool cons uc ion,
we sugges ha bo h sou ces o a ia ion should be aken in o accoun . This s udy demons a es ha DNA pooling
allows as and high- h oughpu de e mina ion o allele equencies in A lan ic salmon enabling cos -e icien
iden i ica ion o in o ma i e ma ke s o disc imina ion o popula ions a a ious geog aphical scales, as well as
iden i ica ion o loci con olling ecologically and economically impo an ai s.
Keywo ds: DNA pooling, A lan ic salmon, SNP, Allele equency es ima ion, Allelo yping, Popula ion genomics
* Co espondence: [email p o ec ed]
2
Depa men o Biology, Di ision o Gene ics and Physiology, Uni e si y o
Tu ku, Tu ku 20014, Finland
Full lis o au ho in o ma ion is a ailable a he end o he a icle
© 2013 Oze o e al.; licensee BioMed Cen al L d. This is an Open Access a icle dis ibu ed unde he e ms o he C ea i e
Commons A ibu ion License (h p://c ea i ecommons.o g/licenses/by/2.0), which pe mi s un es ic ed use, dis ibu ion, and
ep oduc ion in any medium, p o ided he o iginal wo k is p ope ly ci ed.
Oze o e al. BMC Genomics 2013, 14:12
h p://www.biomedcen al.com/1471-2164/14/12
Backg ound
Technological ad ances in polymo phism de ec ion and
geno yping ha e made he single nucleo ide polymo ph-
isms (SNPs) he ma ke o choice o many high densi y
geno yping s udies [1,2]. High- h oughpu mic oa ays
con aining assays o housands o SNPs a e becoming
a ailable o a numbe o non-model o ganisms [1-3], and
being used mo e equen ly in ecological and e olu iona y
s udies, including popula ion gene ics s udies e.g. [4-7],
QTL iden i ica ion e.g. [8], pa en age de e mina ion e.g.
[9-11], and mixed s ock analysis e.g. [12-15].
Despi e he ecen echnical ad ances, geno yping la ge
numbe s o indi iduals wi h housands o SNPs emains
p ohibi i ely expensi e o many esea ch g oups. Fu he -
mo e, many popula ion gene ic s udies a e based on popu-
la ion allele equency a he han indi idual geno ype
da a. The e o e, de e mina ion o allele equencies om
pooled DNA samples, i.e. ‘allelo yping’, has been sugges ed
mo e han 30 yea s ago as a cos -e ec i e al e na i e o in-
di idual geno yping ( e iewed by Sham e al. [16]). Se e al
s udies ha e success ully used his app oach in genome-
wide associa ion s udies ha compa e he allele equen-
cies be ween cases and con ols e.g. [17-23]. These s udies
ha e demons a ed sa is ac o y accu acy and epea abili y,
and he DNA pooling app oach can educe cos s by as
much as 100- old depending on he numbe o samples
[16,21,23].
While he allelo yping o DNA pools can subs an ially e-
duce he cos s compa ed o indi idual sample by sample
geno yping, his app oach is no wi hou disad an ages.
Fi s , a ious sou ces o e o occu du ing he allele e-
quency es ima ion om DNA pools. Acco ding o Ea p
e al. [23], a ia ion in oduced o allele equency es i-
ma es can be di ided in o ou ca ego ies: (i) wi hin a ay;
(ii) be ween a ays; (iii) be ween independen ly cons uc ed
iden ical pools, and (i ) be ween pools cons uc ed om
di e en indi iduals o he same popula ion (biological
eplica es). The e o e, in o de o ob ain eliable allele e-
quency es ima es using DNA pooling i is impo an o
e alua e he magni ude and ela i e impo ance o di e en
sou ces o e o [23,24]. In addi ion, DNA pooling gene -
ally does no p o ide in o ma ion abou haplo ype e-
quency and despi e ecen compu a ional imp o emen s
[25,26] esol ing he phase ambigui y emains a challenge
o la ge numbe o loci [27]. Howe e , despi e he popu-
la i y o DNA pooling in gene ic associa ion s udies, only
ew s udies o da e ha e u ilized allelo yping app oach o
cha ac e ize in e -popula ion a ia ion e.g. [28].
He e, we es ed he use ulness o DNA pooling o a i s
ime using an A lan ic salmon (Salmo sala L.) Illumina
SNP-chip o ob ain accu a e allele equency es ima es o
mul iple A lan ic salmon popula ions and e alua ed he
impo ance o di e en sou ces o e o s a ising om alle-
lo yping. Fi s , we assessed he e ec o DNA pool
cons uc ion and be ween-a ay a ia ions on allele e-
quency es ima es. Subsequen ly, he e ec o clus e sepa -
a ion sco es (pa ame e ha summa izes he sepa a ion o
h ee geno ype classes in he he a dimension), wo al e na-
i e sou ces o he a (a alue be ween 0 and 1 which
de ines he geno ype; 0 = AA, 1 = BB, 0.5 = AB) and DNA
pool size on allele equency es ima ion we e e alua ed.
Finally, wo al e na i e quali y con ol (QC) il e s we e
es ed o selec op imal se s o SNP loci o subsequen
popula ion gene ic analysis.
Resul s and discussion
In o al, 56 A lan ic salmon DNA pools om 14 popula-
ions we e analyzed using an A lan ic salmon SNP-chip
[29,30] ca ying p obes o 5568 SNP ma ke s 3928 o
which we e bi-allelic. A e excluding 1640 non bi-
allelic ma ke s and 31 bi-allelic loci due o low call a e
(< 95%) (see Addi ional ile 1, Figu e S1a) he epea -
abili y o allelo yping om DNA pools was es ed o
3897 loci.
A ay- s. pool-cons uc ion a ia ion
The expe imen al design desc ibed in Table 1 p o ided 56
es ima es o a ay- a ia ion and 52 o pool-cons uc ion
a ia ion in he he a alue. The mean a ay- a ia ion pe
SNP a ied om he a 0.000 o 0.089, whe eas he mean
pool-cons uc ion a ia ion o he a anged om 0.000 o
0.069. The es ima ed a ia ion o he a be ween di e en
a ays (i.e. a ay- a ia ion using iden ical DNA pools) was 20%
highe compa ed o a ia ion a ising om DNA pool con-
s uc ion (median
a ay
=0.012 s.median
pool-cons uc ion
=0.010,
non-pa ame ic Mann–Whi ney U- es , P< 0.0001) (Figu e 1).
These esul s sugges ha i is mo e impo an o conside
a ia ion a ising om di e en a ays han a ia ion asso-
cia ed wi h pool cons uc ion [22-24,31]. This is in line
wi h he ea lie s udies sugges ing ha unning he same
DNA pool in mul iple a ays should be p e e ed o e con-
s uc ion and analysis o mul iple DNA pools wi hin he
same a ay [18,22,32]. Howe e , conside ing he ela i ely
simila le els o a ia ion associa ed wi h he a ay and pool
eplica es, u u e s udies should inco po a e bo h sou ces o
a ia ionin heexpe imen aldesign o eliable es ima ion
o allele equencies om DNA pools.
Es ima ion o allele equencies om DNA pools
The allele equencies o 3631 SNPs ha passed he qual-
i y con ol (see below) we e es ima ed om DNA pools
using e e ence alues o he a p o ided by CIGENE and
e e ence alues o he a de i ed om he geno yping o
106 indi iduals used in pool cons uc ion. Compa ison o
he wo se s o he a alues e ealed a small, bu signi ican ,
di e ence in allele equency es ima es. Using indi idual geno-
ypes om his s udy o de i e e e ence alues o he a p o-
ided sligh ly highe accu acy in allele equency es ima es
Oze o e al. BMC Genomics 2013, 14:12 Page 2 o 9
h p://www.biomedcen al.com/1471-2164/14/12
compa ed o he la ge (n = 300) bu un ela ed da ase p o-
ided by CIGENE (median e o
106
= 0.020 –0.023 s. me-
dian e o
CIGENE
=0.025–0.028; Mann–Whi ney U- es ,
all es s, P< 0.0001; Figu e 2). E o s associa ed wi h
allele equency es ima ions using e e ence alues o
he a om wo di e en sou ces we e signi ican ly
co ela ed (Pea son’s = 0.640 –0.684, P< 0.0001)
sugges ing ha small numbe o SNPs su e om la -
ge e o i espec i e o he sou ce o e e ence alues
o he a while he majo i y o loci ha e ela i ely low
e o a es. Taken oge he , hese esul s sugges ha
e en ela i ely small numbe o indi iduals (~ 100) is
su icien o gene a e eliable e e ence alues o
he a. Howe e , because all h ee geno ype classes a e
needed o accu a e es ima ion o allele equencies,
using ela i ely small numbe o indi iduals esul ed in
loss o SNPs as no all geno ypes we e obse ed in he
e e ence da ase s (3631 s. 3138 SNPs based on
CIGENE and ou da a, espec i ely).
We obse ed e y high conco dance be ween allele e-
quency es ima es de i ed om DNA pools and om indi-
idual geno yping (Pea son’s =0.991–0.992, all es s,
P< 0.0001, Figu e 3). This demons a es he accu acy o
he DNA pooling app oach in A lan ic salmon and is con-
sis en wi h ea lie s udies in o he species using Illumina
bead-a ay pla o m. Fo example, high co ela ion be-
ween allele equency es ima es de i ed om indi idual
geno yping and DNA pools ha e been obse ed in humans
(Pea son’s = 0.969) and ca le (Pea son’s = 0.992 –0.994)
[18,33]. The numbe o indi iduals in he DNA pool had
only a mino e ec on he allele equency es ima ion
(Figu e 3) as he e o be ween ue and es ima ed al-
lele equencies was small and simila o all h ee
pool sizes (median e o = 0.023 –0.025, Figu e 2). The e-
o e, ou esul s sugges ha i is possible o ob ain accu -
a e allele equency es ima es using DNA pools consis ing
o ela i ely small numbe o indi iduals (n ≥35). How-
e e , la ge pool sizes should be always p e e ed o e
Table 1 In o ma ion abou popula ions, hei geog aphic loca ions, numbe o indi iduals and numbe o pool
eplica es s udied
Popula ion Numbe o indi iduals included in he pools and numbe o a ay
and pool cons uc ion eplica es (in b acke s)
Numbe o samples o indi idual geno yping
Pool-size 1 Pool-size 2 Pool-size 3
No wegian Sea coas
Al a 50 (2, 2) 70 (2, 2) 6
Laukhelle 35 (2, 0) 43 (2, 2) 5
Reppa jo del 50 (2, 2) 69 (2, 2)
Ba en s Sea coas
Laksel a 50 (2, 2) 67 (3, 2)
Ves e Jakobsel 50 (2, 2) 70 (2, 2)
Tana B u (Teno) 50 (2, 0) 60 (2, 0)
Ka asjoki (Teno) 50 (2, 2) 70 (2, 2)
Ina ijoki (Teno) 50 (2, 0) 67 (2, 2)
Iesjoki (Teno) 6
Neiden 50 (2, 2) 63 (2, 2)
U a 35 (2, 0) 46 (2, 2)
Ti o ka 50 (2, 2) 70 (2, 2)
Kola 35 (3, 3) 50 (3, 3) 70 (3, 3) 67
Pecho a Unya 6
Whi e Sea coas
Ponoi 50 (2, 2) 70 (2, 2)
Va zuga 50 (2, 2) 70 (2, 2) 6
Onega 6
Bal ic Sea coas
Na a 4
To al pooled 905
To al indi idual 106
Oze o e al. BMC Genomics 2013, 14:12 Page 3 o 9
h p://www.biomedcen al.com/1471-2164/14/12
small ones as small numbe o indi iduals may no be ep-
esen a i e o he whole popula ion.
Quali y con ol
One o he impo an pa ame e s o accu a e de e min-
a ion o geno ypes and subsequen allelo yping is clus e
sepa a ion sco e ha quan i ies he disc imina ion be ween
geno ype clus e s o pa icula SNP (see Addi ional ile 1:
Figu e S1b, c, d). Since he he e ozygous clus e can be
indis inguishable om one o bo h homozygous clus e s
o SNP wi h low clus e sepa a ion sco e, exclusion o loci
demons a ing low clus e sepa a ion sco es has been o en
applied [34,35]. To da e, mos o he s udies ha e used a
clus e sepa a ion sco e cu -o <0.35 o exclude low quali y
SNPs e.g. [36,37]. Based on isual inspec ion o SNP clus-
e s in A lan ic salmon, howe e , cu -o alue o 0.4 was
chosen o e icien ly exclude SNPs showing ambiguous
geno ype classes. This esul ed in selec ion o 3631 ou o
a ay
pool-cons uc ion
Va ia ion o he a
0.08
0.06
0.04
0.02
0.00
Figu e 1 Box-plo showing es ima ed a ay- and pool-cons uc ion a ia ion o he a (Mann–Whi ney U- es , P< 0.0001). Ho izon al line,
g ey squa e, whiske s, open ci cles, and s a s indica e median, 25 h and 75 h qua iles, non-ou lie ange, ou lie s and ex eme ou lie s,
espec i ely.
Figu e 2 Box-plo showing e o in allele equency es ima es calcula ed using he a clus e mean alues p o ided by CIGENE o
ob ained om 106 indi iduals (Mann–Whi ney U- es , all es s, P< 0.0001). Ho izon al line, g ey squa e, whiske s, open ci cles, and s a s
indica e median, 25 h and 75 h qua iles, non-ou lie ange, ou lie s and ex eme ou lie s, espec i ely.
Oze o e al. BMC Genomics 2013, 14:12 Page 4 o 9
h p://www.biomedcen al.com/1471-2164/14/12
3897 ma ke s o subsequen analysis. As expec ed, he e o
in allele equency es ima es o SNPs ha ing clus e sepa -
a ion sco e < 0.4 was highe compa ed o SNPs wi h clus e
sepa a ion sco e > 0.4 (Mann–Whi ney U es , bo h o
a ay and pool eplica es, P< 0.0001) (see Addi ional ile 1:
Figu e S2a, b). Mo eo e , he co ela ion be ween allele
equency es ima es de i ed om h ee DNA pools and
om indi idual geno yping o SNPs demons a ing low
clus e sepa a ion sco es (< 0.4) was lowe han o ma -
ke s wi h clus e sepa a ion sco es > 0.4 (Pea son’s =
0.960 –0.969 s. Pea son’s = 0.991 –0.992). In addi ion,
he es ima ed a ia ion o he a was nega i ely co e-
la ed wi h he clus e sepa a ion sco e bo h o a ay
(Pea son’s =−0.346, P< 0.0001) and pool cons uc ion
(Pea son’s =−0.246, P< 0.0001) eplica es (see Addi ional
ile 1: Figu e S3a, b).
While applica ion o QC il e based on clus e sepa -
a ion excludes SNPs ha ing low quali y geno ypes, i is no
able o emo e all loci showing ela i ely high a ia ion in
allele equency es ima es (see Addi ional 1: Figu e S3a, b).
The e o e, applica ion o addi ional QC il e s, e.g. based
on compa isons be ween ‘ ue’and es ima ed allele e-
quencies o based on combina ion o a ia ion in allele
equency es ima es and he e ozygosi y ha e been sug-
ges ed e.g. [28,36,37].
He e, we es ed wo al e na i e QC il e s (uni o m
and sphe ical cu -o ) ha use he e ozygosi y and a i-
a ion in allele equency es ima es (Figu e 4). This
esul ed selec ion o 2879 s. 2880 loci o uni o m and
sphe ical cu -o , espec i ely (Table 2). Majo i y o loci
(2777) ha passed bo h il e s we e he same (Figu e 4).
Howe e , sphe ical il e ing is expec ed o be mo e use ul
han uni o m cu -o as i e ains la ge p opo ion o poly-
mo phic loci wi h mean allele equency 0.2 –0.8 ac oss
Figu e 3 Sca e plo o es ima ed allele equencies om
indi idual geno yping s. pooled DNA. ‘T ue’allele equencies
om indi idual geno yping o Kola popula ion we e compa ed wi h
es ima ed allele equencies o h ee di e en pool sizes: (a) Kola-35
(n =35, = 0.992), (b) Kola-50 (n = 50; = 0.991) and (c) Kola-70
(n = 70; = 0.992).
0.00
0.01
0.02
0.03
0.04
0.05
0.06
0.07
0.08
0.09
0.0 0.2 0.4 0.6 0.8 1.0
Va ia ion o he a
Mean allele equency
Figu e 4 A plo o mean es ima ed allele equencies ac oss 14
popula ions agains a ay- a ia ion. Solid and dashed lines
indica e he bounda ies o sphe ical and uni o m cu -o s,
espec i ely.
Oze o e al. BMC Genomics 2013, 14:12 Page 5 o 9
h p://www.biomedcen al.com/1471-2164/14/12
popula ions, while uni o m il e inc eases he p opo ion
o less a iable loci (Table 2, Figu e 4, Addi ional 1:
Figu e S4). The e o e, o iden i ica ion o eliable and in-
o ma i e SNPs, applica ion o sphe ical il e is p e e able
o e uni o m since i e ec i ely excludes loci wi h ela-
i ely high e o a e compa ed o he in o ma ion con en .
Conclusions
This s udy es ed he e ec i eness o DNA pooling o
ob ain accu a e allele equency es ima es o la ge num-
be o A lan ic salmon popula ions using an Illumina
SNP-chip. We demons a ed ha pooled DNA app oach
p o ides a eliable, accu a e and cos -e ec i e means o
ob aining genome-wide allele equency es ima es o
mul iple popula ions. We p oposed a no el quali y con-
ol il e based on sphe ical cu -o which enables e i-
cien exclusion o loci showing high e o a e and
mino allele equency close o ze o. Ou esul s indica e
ha e en ela i ely small DNA pools (35 indi iduals)
p o ide accu a e allele equency es ima es o a gi en
sample. Despi e o highe le els o a ia ion associa ed
wi h a ay eplica es compa ed o pool cons uc ion we
sugges ha bo h sou ces o a ia ion should be aken in o
accoun . Taken oge he , his s udy demons a es ha
DNA pooling allows as and high- h oughpu de e min-
a ion o allele equencies in A lan ic salmon enabling
cos -e icien iden i ica ion o in o ma i e ma ke s o dis-
c imina ion o salmon popula ions a a ious geog aphical
scales, as well as iden i ica ion o loci con olling ecologic-
ally and economically impo an ai s. Mo eo e , he
main indings o ou s udy based on A lan ic salmon SNP-
chip we e in line wi h hose obse ed o human SNP-
chips, and hus he echnical app oaches desc ibed he ein
a e encou aging o employing allelo yping app oach in
o he species using Illumina SNP-chips o o he SNP
geno yping sys ems and a ays.
Me hods
DNA samples
In o al, 927 A lan ic salmon indi iduals ep esen ing 19
popula ions om No he n Eu ope we e used o indi id-
ual geno yping and/o cons uc ion o DNA pools (Table 1).
Tissue samples ( in clips) we e collec ed om ju eniles
du ing 2006 –2010 and p ese ed in e hanol. To al gen-
omic DNA was ex ac ed acco ding o Elphins one e al.
[38] o using Qiagen DNeasy 96 Blood & Tissue ki s
(Qiagen™) ollowing manu ac u e ’s ecommenda ions.
Quali y con ol o DNA ex ac s
P io o pool cons uc ion, quali y con ol o indi idual
DNA ex ac s was pe o med in wo s eps. Fi s , samples
we e examined o deg ada ion by isual inspec ion on
1% aga ose gels. Samples con aining low molecula
weigh DNA (indica i e o deg ada ion) we e excluded
om u he analysis. Each ex ac was hen es ed o
con amina ion ( he p esence o DNA om mul iple indi-
iduals) by sc eening indi idual samples using 18 mic o-
sa elli e loci [39]; V. Wenne ik, unpublished da a] and
only non-con amina ed A lan ic salmon samples we e
selec ed o u he analysis.
Cons uc ion o DNA pools and SNP geno yping
In o al, 56 DNA pools we e cons uc ed using indi i-
duals om 14 A lan ic salmon popula ions (Table 1).
The adjus men o DNA concen a ion was ca ied ou
in wo s eps. The ini ial concen a ion o DNA samples
was i s adjus ed o 20 ng/μl, measu ed in duplica e
wi h he NanoD op™1000 (The mo Scien i ic) and sub-
sequen ly dilu ed o 10 ng/ul. Indi idual DNA samples
we e pooled (50 ng pe indi idual) and subsequen ly
concen a ed using a DNA concen a o Eppendo
5301. The inal concen a ion o he pools was adjus ed
o 50 ng/μl. Cons uc ed DNA pools we e analyzed
using an A lan ic salmon Illumina SNP-chip [29,30] a
he Cen e o In eg a i e Gene ics (CIGENE), No way.
In addi ion, 106 salmon samples used in pool cons uc-
ion we e geno yped indi idually o guide clus e posi-
ioning and o ob ain he ‘ ue’allele equency o each
locus o he popula ion om he Ri e Kola (Table 1).
Quali y con ol
Geno yping o he 106 indi idual samples was pe o med
using Geno yping module . 1.9.4 (Genome S udio so wa e
. 2011.1, Illumina Inc.), only hose samples wi h > 97%
call a es we e included when calcula ing ‘ ue’allele e-
quencies. SNPs wi h call a es < 95% (i.e. he p opo ion
o indi idual samples success ully geno yped in a locus)
we e elimina ed om he da a se . Th esholds o quali y
con ol (QC) il e ing we e de e mined as in Mu ay
e al. [37] and o es ima ion o allele equencies om
DNA pools, SNPs wi h clus e sepa a ion sco es ≤0.4
we e excluded.
Es ima ion o allele equencies in a pooled DNA samples
In Illumina geno yping, he geno ype is assigned a e con-
e ing aw colo signal da a in o a he a alue which anges
om 0 o 1 and e lec s he ela i e signal con ibu ion
o he 2 al e na e alleles. In heo y, an indi idual
Table 2 Numbe o loci e ained a e applying sphe ical
o uni o m QC il e ing o 3631 SNPs
Fil e Mean allele equency ac oss 14 popula ions
< 0.1 0.1 –0.4 0.4 –0.6 0.6 –0.9 > 0.9 To al
Be o e il e ing 326 1110 845 1044 306 3631
Sphe ical 275 877 690 787 251 2880
Uni o m 308 860 639 785 287 2879
Oze o e al. BMC Genomics 2013, 14:12 Page 6 o 9
h p://www.biomedcen al.com/1471-2164/14/12
homozygous o he B allele would ha e a he a alue close
o 1, an indi idual homozygous o he A allele a alue close
o 0 and a alue o 0.5 would indica e a he e ozygous geno-
ype. Howe e , in eali y a SNP’s he a o geno ype clus-
e s (AA, AB and BB) may a y om 0, 0.5 and 1,
he e o e o es ima ion o allele equency in a pooled
sample, he he a alue o each SNP is compa ed o he
mean he a alues o AA, AB and BB geno ypes calcu-
la ed by geno yping indi idual samples, i.e. he allele e-
quency o he DNA pool can be de i ed by applying
co ec ion algo i hms om compa ing pool-speci ic
alue o he a wi h he e e ence alues o he a om in-
di idual geno yping da a e.g. [40,41].
To ob ain he allele equency es ima e o allele B in
he pool B
pool
Sample posi ion o each pool along he
axis o no malized he a alues we e compa ed o he
e e ence alues o AA, AB and BB geno ype clus e
posi ions o each SNP ( e e ence alues o he a) as in
Janicki & Liu [41].
The ollowing equa ions we e applied [41]:
i θpool ≤θAA; hen Bpool ¼0o
i θAA <θpool <θAB hen Bpool
¼0:5θpool θAA
=θAB θAA
ðÞo
i θAB <θpool <θBB; hen Bpool
¼0:5þ0:5θpool θAB
=θAB θAB
ðÞo
i θpool ≥θBB; hen Bpool ¼1;whe e
θ
pool
is he sample posi ion and θ
AA
,θ
AB
,θ
BB
a e
means o he clus e posi ions o he co esponding
e e ence geno ypes along he axis o no malized he a
alues. The equency o allele A was calcula ed as
A
pool
=1–B
pool
.
Re e ence alues o AA, AB and BB geno ype posi ions
along he axis o no malized he a alues we e ob ained
om indi idual geno yping o 300 A lan ic salmon speci-
mens geno yped in p e ious s udies by CIGENE. As his
da a did no include samples om all he popula ions used
o cons uc he DNA pools, he mean clus e posi ion
alues we e also de i ed om he geno ype classes o 106
indi iduals o igina ing om 8 popula ions ac oss he s udy
a ea (Ri e s: Al a, Laukhelle, Iesjoki, Kola, Va zuga, Onega,
Pecho a Unya and Na a). Fo subsequen analyses,
howe e , e e ence alues o he a p o ided by CIGENE
we e used.
The accu acy o allele equency es ima es was quan i ied
as an absolu e di e ence be ween allele equencies de i ed
om indi idual geno ypes ( e e ed o as ‘ ue’) and allele
equencies es ima ed om DNA pools om he Ri e
Kola popula ion (35, 50 and 70 indi iduals pe pool).
Es ima ion o a ay- and pool-cons uc ion a ia ion
To es ima e he wi hin-pool a ia ion o he a, eplica es o
he same DNA pool we e un on di e en a ays (a ay
eplica es, as in Ea p e al. [23]) (Table 1). To assess he
a ia ion in he a alues in oduced by pool cons uc ion,
independen ly cons uc ed pools consis ing he same DNA
ex ac s we e un on same a ay (pool cons uc ion epli-
ca es, as in Ea p e al. [23]) (Table 1). To e alua e he e ec
o numbe o indi iduals in he DNA pool on allele e-
quency es ima ion, DNA pools wi h a ying numbe o in-
di idual DNA ex ac s we e cons uc ed (Table 1).
Va ia ion o he a wi hin a SNP locus was es ima ed simi-
la o Macg ego [31]. The a ay- a ia ion was calcula ed as
he mean di e ence o all possible pai -wise compa isons o
he a alues among echnical eplica es o he same pool
allelo yped on di e en a ays. The pool-cons uc ion a i-
a ion was calcula ed as he mean di e ence o all possible
pai -wise compa isons o he a alues among echnical
eplica es o he independen ly cons uc ed DNA pools
con aining same indi iduals allelo yped on he same a ay.
Addi ional ile
Addi ional ile 1: Figu e S1. Example o SNP loci ailed o pass QC:
a) call a e < 95%; b) clus e sepa a ion < 0.40; and SNP loci me QC
equi emen s: c) clus e sepa a ion = 0.41 and d) call a e 100%, clus e
sepa a ion = 1.00. Figu e S2. Box-plo showing es ima ed a ia ion o
he a in wo se s o SNPs wi h clus e sepa a ion sco e < 0.4 and > 0.4
o (a) a ay and (b) pool cons uc ion eplica es (bo h es s, Mann–
Whi ney U es , P< 0.0001). Ho izon al line, g ey squa e, whiske s, open
ci cles, and s a s indica e median, 25 h and 75 h qua iles, non-ou lie
ange, ou lie s and ex eme ou lie s, espec i ely. Figu e S3. A signi ican
nega i e co ela ion be ween (a) a ay-(Pea son’s =−0.346, P< 0.0001)
and (b) pool-cons uc ion (Pea son’s =−0.246, P< 0.0001) a ia ion and
clus e sepa a ion sco es. Figu e S4. P opo ion o loci emained in each
allele equency class a e applica ion o (a) uni o m and
(b) sphe ical il e .
Compe ing in e es s
The au ho s decla e ha hey ha e no compe ing in e es s.
Au ho s’con ibu ions
M.O. cons uc ed he DNA pools, pe o med he da a analysis and lead in
d a ing he manusc ip . J.-P.V. and A.V. designed he s udy and signi ican ly
con ibu ed o he da a analysis and he w i ing o he manusc ip . V.W. ook
pa in designing he s udy, p o ided mic osa elli e da a and oge he wi h
E.N. and S.P. collec ed biological samples. M.K. pe o med SNP a ay
sc eening and ini ial da a analysis. All au ho s ead and app o ed he inal
manusc ip .
Acknowledgemen s
We hank Rogelio Diaz-Fe nandez and K is iina Haapanen o labo a o y
assis ance and wo anonymous e iewe s o hei help ul commen s in
imp o ing his manusc ip . This s udy was unded by he Eu opean Union,
Kola c ic ENPI CBC p ojec KO197 (M.O., E.N. and S.P), Academy o Finland
(J.-P.V.), No wegian Di ec o a e o Na u e Managemen (V.W.), No wegian
Resea ch Council (V.W.) and Es onian Science Founda ion (g an numbe s
6802, 8215 o A.V.). This publica ion has been p oduced wi h he assis ance
o he Eu opean Union, bu he con en s can in no way be aken o e lec
he iews o he Eu opean Union.
Oze o e al. BMC Genomics 2013, 14:12 Page 7 o 9
h p://www.biomedcen al.com/1471-2164/14/12
Au ho de ails
1
Ke o Suba c ic Resea ch Ins i u e, Uni e si y o Tu ku, Tu ku 20014, Finland.
2
Depa men o Biology, Di ision o Gene ics and Physiology, Uni e si y o
Tu ku, Tu ku 20014, Finland.
3
Ins i u e o Ma ine Resea ch, PO Box 1870,
No dnes N-5817, Be gen, No way.
4
Finnish Game and Fishe ies Resea ch
Ins i u e, Raken ajan ie 3,PL 413, 90014, Oulun yliopis o, Finland.
5
F eshwa e
Resou ces Labo a o y, Knipo i ch Pola Resea ch Ins i u e o Ma ine Fishe ies
and Oceanog aphy, 6. Knipo i ch S ee , 183767, Mu mansk, Russia.
6
Cen e
o In eg a i e Gene ics (CIGENE), Depa men o Animal and Aquacul u al
Sciences (IHA), No wegian Uni e si y o Li e Sciences, PO Box 5003, Ås,
No way.
7
Depa men o Aquacul u e, Ins i u e o Ve e ina y Medicine and
Animal Science, Es onian Uni e si y o Li e Sciences, 51014, Ta u, Es onia.
Recei ed: 27 June 2012 Accep ed: 2 Janua y 2013
Published: 16 Janua y 2013
Re e ences
1. Ga in MR, Sai oh K, Gha e AJ: Applica ion o single nucleo ide
polymo phisms o non-model species: a echnical e iew. Mol Ecol Res
2010, 10:915–934.
2. Seeb JE, Ca alho G, Hause L, Naish K, Robe s S, Seeb LW: Single-
nucleo ide polymo phism (SNP) disco e y and applica ions o SNP
geno yping in nonmodel o ganisms. Mol Ecol Res 2011, 11:1–8.
3. Coa es BS, Sume o d DV, Mille NJ, Kim KS, Sapping on TW, Sieg ied BD,
Lewis LC: Compa a i e pe o mance o single nucleo ide polymo phism
and mic osa elli e ma ke s o popula ion gene ic analysis. J He ed 2009,
100:556–564.
4. O’Malley KG, Cama a MD, Banks MA: Candida e loci e eal gene ic
a ia ion be ween empo ally di e gen mig a o y uns o Chinook
salmon (Onco hynchus shawy scha). Mol Ecol 2007, 16:4930–4941.
5. Tenesa A, Na a o P, Hayes BJ, Du y DL, Cla ke GM, Godda d ME, Vissche
PM: Recen human e ec i e popula ion size es ima ed om linkage
disequilib ium. Genome Res 2007, 17:520–526.
6. Zenge KR, Kha ka MS, Ca anagh JAL, Hawken RJ, Raadsma HW: Genome-
wide gene ic di e si y o Hols ein F iesian ca le e eals new insigh s
in o Aus alian and global popula ion a iabili y, including impac o
selec ion. Anim Gene 2007, 38:7–14.
7. Kelle I, Vel sos P, Nichols RA: The equency o DNA a ian s wi hin
indi iduals p o ides e idence o popula ion his o y and gene low
ac oss a g asshoppe hyb id zone. E olu ion 2008, 62:833–844.
8. Boulding EG, Culling M, Glebe B, Be g PR, Lien S, Moen T: Conse a ion
genomics o A lan ic salmon: SNPs associa ed wi h QTLs o adap i e
ai s in pa om ou ans-A lan ic backc osses. He edi y 2008,
101:381–391.
9. Hea on MP, Ha hay GP, Benne GL, S one RT, G osse WM, Casas E, Keele
JW, Smi h TP, Chi ko-McKown CG, Laeg eid WW: Selec ion and use o SNP
ma ke s o animal iden i ica ion and pa e ni y analysis in U.S. bee
ca le. Mamm Genome 2002, 13:272–281.
10. Ande son EC, Ga za JC: The powe o single nucleo ide polymo phisms
o la ge-scale pa en age in e ence. Gene ics 2006, 172:2567–2582.
11. Toka ska M, Ma shall T, Kowalczyk R, Wójcik JM, Pe oldi C, K is ensen TN,
Loeschcke V, G ege sen VR, Bendixen C: E ec i eness o mic osa elli e and
SNP ma ke s o pa en age and iden i y analysis in species wi h low
gene ic di e si y: he case o Eu opean bison. He edi y 2009, 103:326–332.
12. Smi h C, Templin W, Seeb J, Seeb L: Single Nucleo ide Polymo phisms
(SNPs) p o ide apid and accu a e es ima es o he p opo ions o U.S.
and Canadian Chinook salmon caugh in Yukon Ri e ishe ies. NAmJ
Fishe Man 2005, 25:944–953.
13. Na um SR, Banks M, Beacham TD, Bellinge MR, Campbell MR, Dekoning J,
Elz A, Gu h ie CM 3 d, Koz kay C, Mille KM, Mo an P, Phillips R, Seeb LW,
Smi h CT, Wa hei K, Young SF, Ga za JC: Di e en ia ing salmon
popula ions a b oad and ine geog aphical scales wi h mic osa elli es
and single nucleo ide polymo phisms. Mol Ecol 2008, 17:3464–3477.
14. Hess JE, Ma ala AP, Na um SR: Compa ison o SNPs and mic osa elli es o
ine-scale applica ion o gene ic s ock iden i ica ion o Chinook salmon
in he Columbia Ri e Basin. Mol Ecol Res 2011, 11(Suppl 1):137–149.
15. Ka lsson S, Moen T, Lien S, Glo e KA, Hinda K: Gene ic gene ic di e ences
be ween a med and wild A lan ic salmon iden i ied om a 7K
SNP-a ay. Mol Ecol Resou 2011, 11(Suppl 1):247–253.
16. Sham P, Bade JS, C aig I, O'Dono an M, Owen M: DNA Pooling: a ool o
la ge-scale associa ion s udies. Na Re Gene 2002, 3:862–871.
17. S okowski RP, Pan PV, Dadd T, Fe eday A, Hinds DA, Ja man C, Filsell W,
Ginge RS, G een MR, an de Oude aa FJ, Cox DR: A genomewide
associa ion s udy o skin pigmen a ion in a Sou h Asian popula ion. Am
J Hum Gene 2007, 81:1119–1132.
18. Ab aham R, Mosk ina V, Sims R, Hollingwo h P, Mo gan A, Geo gie a L,
Dowzell K, Cichon S, Hillme AM, O’Dono an MC, Williams J, Owen MJ, Ki o
G: A genome-wide associa ion s udy o la e-onse Alzheime ’s disease
using DNA pool cons uc ion. BMC Med Genomics 2008, 1:44.
19. B own KM, Macg ego S, Mon gome y GW, C aig DW, Zhao ZZ, Iyadu ai K,
Hende s AK, Home N, Campbell MJ, S a k M, Thomas S, Schmid H, Holland
EA, Gillande s EM, Du y DL, Maskiell JA, Je ann J, Fe guson M, S ephan DA,
Cus AE, Whi eman D, G een A, Olsson H, Puig S, Ghio zo P, Hansson J,
Demenais F, Golds ein AM, G uis NA, Elde DE, Bishop JN, Ke o d RF, Giles
GG, A ms ong BK, Ai ken JF, Hoppe JL, Ma in NG, T en JM, Mann GJ,
Haywa d NK: Common sequence a ian s on 20q11.22 con e melanoma
suscep ibili y. Na Gene 2008, 40:838–840.
20. Comabella M, C aig DW, Camina-Ta o M, Mo cillo C, Lopez C, Na a o A, Rio
J, Bioma ke MS S udy G oup, Mon alban X, Ma in R: Iden i ica ion o a
no el isk locus o mul iple scle osis a 13q31.3 by a pooled
genomewide scan o 500,000 single nucleo ide polymo phisms.
PLoS One 2008, 3:e3490.
21. Macg ego S, Zhao ZZ, Hende s A, Nicholas MG, Mon gome y GW, Vissche
PM: Highly cos -e icien genome-wide associa ion s udies using DNA
pools and dense SNP a ays. Nucleic Acids Res 2008, 36:e35.
22. Huang Y, Hinds DA, Lihong Q, P en ice RL: Pooled e sus indi idual
geno yping in a b eas cance genome-wide associa ion s udy.
Gene Epidemiol 2010, 34:603–612.
23. Ea p MA, Rahmani M, Chew K, B ook-Wilson A: Es ima es o a ay and pool
cons uc ion a ia ion o planning e icien DNA-pool cons uc ion
genome wide associa ion s udies. BMC Med Genomics 2011, 4:81.
24. Macg ego S, Vissche PM, Mongome y G: Analysis o pooled DNA samples
on high densi y a ays wi hou p io knowledge o di e en ial
hyb idiza ion a es. Nucleic Acid Resea ch 2006, 34:e55.
25. Ki kpa ick B, A menda iz CS, Ka p RM, Halpe in E: HAPLOPOOL: imp o ing
haplo ype equency es ima ion h ough DNA pools and phylogene ic
modeling. Bioin o ma ics 2007, 23:3048–3055.
26. Zhang H, Yang H-C, Yang Y: PoooL: an e icien me hod o es ima ing
haplo ype equencies om la ge DNA pools. Bioin o ma ics 2008,
24:1942–1948.
27. Kuk AYC, Xu J, Yang Y: A s udy o he e iciency o pooling in haplo ype
es ima ion. Bioin o ma ics 2010, 26:2556–2563.
28. Chiang CW, Gajdos ZK, Ko n JM, Ku u illa FG, Bu le JL, Hacke R, Guiducci
C, Nguyen TT, Wilks R, Fo es e T, Haiman CA, Hende son KD, Le Ma chand
L, Hende son BE, Palme MR, McKenzie CA, Lyon HN, Coope RS, Zhu X,
Hi schho n JN: Rapid assessmen o gene ic ances y in popula ions o
unknown o igin by genome-wide geno yping o pooled samples. PLoS
Gene 2010, 6:1–11.
29. Bou e V, Ken MP, P imme CR, Vasemägi A, Ka lsson S, Hinda K,
McGinni y P, Ve spoo E, Be na chez L, Lien S: SNP-a ay e eals genome
wide pa e ns o geog aphical and po en ial adap i e di e gence ac oss
he na u al ange o A lan ic salmon (Salmo sala ). Mol Ecol. doi:10.1111/
mec.12003.
30. Lien S, Gidskehaug L, Moen T, Hayes BJ, Be g PR, Da idson WS, Omhol SW,
Ken MP: A dense SNP-based linkage map o A lan ic salmon (Salmo sala )
e eals ex ended ch omosome homeologies and s iking di e ences in
sex-speci ic ecombina ion pa e ns. BMC Genomics 2011, 12:615.
31. Macg ego S: Mos pool cons uc ion a ia ion is a ay-based DNA pool
cons uc ion is a ibu able o a ay e o han pool cons uc ion e o .
Eu J Hum Gene 2007, 15:501–504.
32. Shi man S, Bhom a A, Smiley S, W ay NR, James MR, Ma in NG, He ema
JM, An SS, Neale MC, an de Oo d EJCG, Kendle KS, Chen X, Boomsma DI,
Middeldo p CM, Ho enga JJ, Slagboom PE, Flin J: A whole genome
associa ion s udy o neu o icism using DNA pool cons uc ion. Mol
Psychia 2008, 13:302–312.
33. Uemo o Y, Sasago N, Abe T, Okada H, Ma uoka H, Nakajima H, Shoji N,
Ma uyama S, Kobayashi N, Mannen H, Kobayashi E: P ac ical capabili y o a
DNA pool-based genome-wide associa ion s udy using Bo ineSNP50
a ay in a ca le popula ion. Anim Sci J 2012. doi:10.1111/j.1740-
0929.2012.01022.x.
34. Hy en D, Song Q, Choi I-Y, Yoon M-S, Spech JE, Ma ukumalli LK, Nelson RL,
Shoemake RC, Young ND, C eagn PB: High- h oughpu geno yping wi h
Oze o e al. BMC Genomics 2013, 14:12 Page 8 o 9
h p://www.biomedcen al.com/1471-2164/14/12
he GoldenGa e assay in he complex genome o soybean. Theo Appl
Gene 2008, 116:945–952.
35. Leppoi e in C, F ige io J-M, Ga nie -Gé é P, Salin F, Ce e a M-T, Vo nam B,
Ha eng L, Plomion C: In i o s. in silico de ec ed SNPs o he
de elopmen o a geno yping a ay: wha can we lea n om a
non-model species? PLosOne 2010, 5:11034.
36. Kwee LC, Liu Y, Haynes C, Gibson JR, S one A, Schichman SA, Kamel F,
Nelson LM, Topol B, Van Den Eeden SK, Tanne CM, Cudkowicz ME, G asso
DL, Lawson R, Mu alidha S, Oddone EZ, Schmid S, Hause MA: A high-
densi y genome-wide associa ion sc een o spo adic ALS in US e e ans.
PLoS One 2012, 7:e32768.
37. Mu ay SS, Smi h EN, Villa asa N, Nahey T, Lande J, Goldbe g H, Shaw M,
Rosen hal L, Ramza B, Alaeddini J, Han X, Damani S, Soykan O, Kowal RC,
Topol EJ, GAME In es iga o s: Genome-wide associa ion o implan able
ca dio e e -de ib illa o ac i a ion wi h li e- h ea ening a hy hmias.
PLoS One 2012, 7:e25387.
38. Elphins one MS, Hin en GN, Ande son MJ, Nock CJ: An inexpensi e and
high- h oughpu p ocedu e o ex ac and pu i y o al genomic DNA o
popula ion s udies. Mol Ecol No es 2003, 3:317–320.
39. Vähä J-P, E kina o J, Niemelä E, Saloniemi I, P imme CR, Johansen M,
S enning M, B ø s S: Tempo ally s able popula ion-speci ic di e ences in
un iming o one-sea-win e A lan ic salmon e u ning o a la ge i e
sys em. E ol Appl 2011, 4:39–53.
40. Wilkening S, Chen B, Wi enbe ge M, Bu winkel B, Fö s i A, Hemminki K,
Canzian F: Allelo yping o pooled DNA wi h 250 K SNP mic oa ays.
BMC Genomics 2007, 8:77.
41. Janicki P, Liu J: Accu acy o allele equency es ima es in pool DNA analyzes by
high-densi y Illumina Human 610-Quad mic oa ay. P o eomics: In e ne J
Genom; 2009:5.
doi:10.1186/1471-2164-14-12
Ci e his a icle as: Oze o e al.:Cos -e ec i e genome-wide es ima ion
o allele equencies om pooled DNA in A lan ic salmon (Salmo sala
L.). BMC Genomics 2013 14:12.
Submi you nex manusc ip o BioMed Cen al
and ake ull ad an age o :
• Con enien online submission
• Tho ough pee e iew
• No space cons ain s o colo figu e cha ges
• Immedia e publica ion on accep ance
• Inclusion in PubMed, CAS, Scopus and Google Schola
• Resea ch which is eely a ailable o edis ibu ion
Submi you manusc ip a
www.biomedcen al.com/submi
Oze o e al. BMC Genomics 2013, 14:12 Page 9 o 9
h p://www.biomedcen al.com/1471-2164/14/12