Da a Desc ip o : A longi udinal
plasma lipidomics da ase om
child en who de eloped isle
au oimmuni y and ype 1diabe es
San osh Lamichhane
1
, Linda Ahonen
2
, Thomas Spa hol Dy lund
2
, Heli Siljande
3,4
,
Heikki Hyö y
5,6
, Jo ma Ilonen
7,8
, Jo ma Toppa i
9,10
, Rii a Veijola
11,12,13
,
Tuulia Hyö yläinen
14
, Mikael Knip
3,4,15,16
& Ma ej O ešič
1,17
Ea ly p edic ion and p e en ion o ype 1diabe es (T1D) a e cu en ly unme medical needs. P e ious
me abolomics s udies sugges ha child en who de elop T1D a e cha ac e ised by a dis inc me abolic
p ofile al eady de ec able du ing in ancy, p io o he onse o isle au oimmuni y. Howe e , he specifici y
o pe sis en me abolic dis u bances in ela ion T1D de elopmen has no ye been es ablished. He e, we
epo a longi udinal plasma lipidomics da ase om (1)40 child en who p og essed o T1D du ing ollow-
up, (2)40 child en who de eloped single isle au oan ibody bu did no de elop T1D and (3)40 ma ched
con ols (6 ime poin s: 3,6,12,18,24 and 36 mon hs o age). This da ase may help o he esea che s in
s udying age-dependen p og ession o isle au oimmuni y and T1D as well as o he age-dependence o
lipidomic p ofiles in gene al. Al e na i ely, his da ase could mo e b oadly used o he de elopmen o
me hods o he analysis o longi udinal mul i a ia e da a.
Design Type(s) disease s a e design •obse a ional design •longi udinal s udy design
Measu emen Type(s) lipid
Technology Type(s) liquid ch oma og aphy- andem mass spec ome y
Fac o Type(s) age •disease s age •au oan ibody measu emen
Sample Cha ac e is ic(s) Homo sapiens •blood plasma
1
Tu ku Cen e o Bio echnology, Uni e si y o Tu ku and Åbo Akademi Uni e si y, Tu ku 20520, Finland.
2
S eno
Diabe es Cen e Copenhagen, 2820 Gen o e, Denma k.
3
Child en’s Hospi al, Uni e si y o Helsinki and Helsinki
Uni e si y Hospi al, 00290 Helsinki, Finland.
4
Resea ch P og am Uni , Diabe es and Obesi y, Uni e si y o
Helsinki, 00290 Helsinki, Finland.
5
Facul y o Medicine and Li e Sciences, Uni e si y o Tampe e, Tampe e,
Finland.
6
Fimlab Labo a o ies, Pi kanmaa Hospi al Dis ic , Tampe e, Finland.
7
Immunogene ics Labo a o y,
Ins i u e o Biomedicine, Uni e si y o Tu ku, Tu ku, Finland.
8
Clinical Mic obiology, Tu ku Uni e si y Hospi al,
Tu ku, Finland.
9
Ins i u e o Biomedicine, Cen e o In eg a i e Physiology and Pha macology, Uni e si y o
Tu ku, Tu ku, Finland.
10
Depa men o Pedia ics, Tu ku Uni e si y Hospi al, Tu ku, Finland.
11
Depa men o
Paedia ics, PEDEGO Resea ch Uni , Medical Resea ch Cen e, Uni e si y o Oulu, Oulu, Finland.
12
Depa men o
Child en and Adolescen s, Oulu Uni e si y Hospi al, Oulu, Finland.
13
Depa men o Women’s and Child en’s
Heal h, Ka olinska Ins i u e , S ockholm, Sweden.
14
Depa men o Chemis y, Ö eb o Uni e si y, 702 81 Ö eb o,
Sweden.
15
Tampe e Cen e o Child Heal h Resea ch, Tampe e Uni e si y Hospi al, Tampe e, Finland.
16
Folkhälsan Resea ch Cen e , Helsinki, Finland.
17
School o Medical Sciences, Ö eb o Uni e si y, 702 81 Ö eb o,
Sweden. Co espondence and eques s o ma e ials should be add essed o M.K. (email: mikael.knip@helsinki.fi)
o M.O. (email: ma ej.o esic@u u.fi)
OPEN
Recei ed: 1June 2018
Accep ed: 26 Sep embe 2018
Published: 13 No embe 2018
www.na u e.com/scien i icda a
SCIENTIFIC DATA |5:180250 |DOI: 10.1038/sda a.2018.250 1
Backg ound & Summa y
Type 1 diabe es (T1D) is a ch onic, au oimmune disease caused by p og essi e loss o insulin-sec e ing
capaci y due o he selec i e dea h o be a cells in he isle s o Lange hans, o which he e a e mo e han
one million in he human panc eas
1
. The age o onse is usually be ween 5–15 yea s, bu , in ecen yea s,
many child en be o e 5 yea s o age a e being a ec ed
2
. Al hough app oxima ely 80% o subjec s wi h
T1D ca y defined isk-associa ed geno ypes a he human leukocy e an igen (HLA) locus, only 3–7% o
he ca ie s o such gene ic isk ma ke s go on o de elop o e disease. Se ocon e sion o isle
au oan ibody posi i i y is he fi s de ec able signal demons a ing ini ia ion o au oimmuni y and isk o
p og ession owa ds diabe es
3,4
. Howe e , whils se ocon e sion o au oan ibody posi i i y p ecedes
clinical disease by mon hs o yea s, he poin a which se ocon e sion occu s may al eady be oo la e o
he apeu ic app oaches aimed a p e en ing p og ession o o e diabe es.
P e ious me abolomic s udies sugges ha child en who de elop isle au oimmuni y and/o T1D a e
cha ac e ised by specific me abolic dis u bances p io o he fi s appea ance o isle au oan ibodies
5–7
.
Me abolic p ofiling may hus be o clinical ele ance by p o iding a complimen a y ool o es ima ing
isk o p og ession o T1D. Howe e , he unde lying causes o hese ea ly me abolic dis u bances and
hei link o disease p og ession a e s ill la gely unknown. I is also no ye es ablished i he obse ed,
pe sis en me abolic changes a e specifically associa ed wi h he p og ession o T1D, o mo e b oadly, i
hey a e associa ed wi h p og ession o isle au oimmuni y and i espec i e o disease ou come.
He e we p o ide a longi udinal plasma lipidomics da ase ob ained om child en pa icipa ing in a
p ospec i e, bi h coho Type 1 Diabe es P edic ion and P e en ion s udy (DIPP). Th ee s udy
g oups we e examined: child en who (1) p og essed o T1D (PT1D), (2) de eloped a leas a single
isle au oan ibody (Ab) du ing ollow-up bu did no p og ess o T1D (P1Ab), and (3) con ols (CTR)
who emained au oan ibody nega i e and heal hy du ing he ollow-up un il 15 yea s o age. We
analysed 428 plasma samples om 120 child en (40 PT1D, 40 P1Ab and 40 CTR). The samples we e
collec ed up o six di e en ime poin s co esponding o he ages o 3, 6, 12, 18, 24, and 36 (o abo e)
mon hs (Fig. 1). These age g oups we e selec ed wi h he objec i e o unde s anding he changes in
lipidomic p ofile p eceding p io o o e T1D. We pe o med un a ge ed lipidomics using ul a-high-
pe o mance liquid ch oma og aphy combined wi h quad upole ime-o -fligh mass spec ome y
(UHPLC-QTOF-MS). Bo h aw and p e-p ocessed da ase s we e deposi ed in he Me aboLigh s
eposi o y (Da a Ci a ion 1). Along wi h he a o emen ioned lipidomic da a, we p o ide in o ma ion
on he ype o isle au oan ibodies obse ed in his longi udinal se ing. In his s udy, sphingomyelins
(SMs) we e ound o be pe sis en ly down egula ed in PT1D as compa ed o he P1Ab and CTR
g oups
8
. T iacylglyce ols (TGs) and phospha idylcholines (PCs) we e mainly down egula ed in PT1D
as compa ed o P1Ab a he age o 3 mon hs. These esul s sugges ha dis inc lipidomic signa u es
cha ac e ise child en who p og essed o isle au oimmuni y o clinical T1D. Lipidomic p ofiling may
hus be help ul in he iden ifica ion o a - isk child en be o e he ini ia ion o au oimmuni y.
This da a desc ip o is one o he fi s longi udinal lipidomic da ase s allowing he in es iga ion o
p og ession o isle au oimmuni y/T1D du ing he ea ly p od omal phases o disease de elopmen .
Conside ing he longi udinal s udy design, his clinical da ase may ha e many uses. Fi s ly, i may assis
o he esea che s in hei s udies o ea ly pa hogenesis o T1D and, po en ially, o he immune-media ed
inflamma o y diseases. I may also allow o he esea che s o s udy he age-dependen p og ession o
lipidomic p ofiles du ing in ancy. Finally, he da ase has g ea po en ial o be used in he de elopmen
and es ing o algo i hms o he analysis o mul i a ia e longi udinal/p ospec i e da a.
Me hods
These me hods a e expanded e sions o desc ip ions in ou ela ed wo k
8
.
S udy design
The plasma samples o lipidomics analysis we e ob ained om he Finnish Type 1 Diabe es P e en ion
and P edic ion (DIPP) s udy
9
. The DIPP s udy has sc eened mo e han 220,000 newbo n in an s
o HLA-con e ed suscep ibili y o T1D a h ee uni e si y hospi als in Finland: Tu ku, Tampe e and
Oulu
10
. O e 25,000 in an s we e iden ified as ha ing an inc eased gene ic isk and app oxima ely 17,000
amilies joined he ollow-up s udy, which in ol es egula s udy cen e isi s (wi h an in e al o 3–6
mon hs). The E hics and Resea ch Commi ee o he pa icipa ing Uni e si ies and Hospi als app o ed
he s udy p o ocol. The s udy was conduc ed acco ding o he guidelines in he Decla a ion o Helsinki.
All amilies p o ided w i en in o med consen o pa icipa ion in he s udy. A e e y isi , blood
samples we e collec ed o measu e he i e o T1D-associa ed isle au oan ibodies. Non- as ing blood
samples we e collec ed in o sodium ci a e ubes. Plasma was sepa a ed wi hin 30 min o collec ion by
cen i uga ion a 1600g o 20 min a oom empe a u e, aliquo ed, and s o ed a −80 °C un il analysed.
Now, 1,663 o hose child en (9.8%) ha e se ocon e ed o posi i i y o one au oan ibody du ing he
ollow-up, 808 (4.8%) ha e de eloped mul iple au oan ibodies, and 510 (3.1%) ha e p og essed o
clinical T1D.
The 120 in an s included in he cu en da a desc ip o we e selec ed om a subse o DIPP child en
om he ci y o Tampe e. Up o six longi udinal samples pe child we e collec ed be ween 1998 and 2012,
co esponding o he ages o 3, 6, 12, 18, 24 and 36 mon hs (Fig. 1, Table 1). The de ails o he ages and
selec ed cha ac e is ics o he cu en s udy subjec s a e gi en in he me ada a (Da a Ci a ion 1). The h ee
www.na u e.com/sda a/
SCIENTIFIC DATA |5:180250 |DOI: 10.1038/sda a.2018.250 2
s udy g oups we e ma ched by HLA-associa ed diabe es isk, gende and he pe iod o bi h. In o al, 428
plasma samples we e selec ed and analysed o his s udy.
HLA geno yping
HLA-con e ed suscep ibili y o T1D was analysed using co d blood samples as p e iously desc ibed
11
.
B iefly, he HLA-geno yping was pe o med wi h a ime- esol ed fluo ome y-based assay o ou
alleles using lan hanide chela e labeled, sequence-specific oligonucleo ide p obes de ec ing he DQB1
∗
02,
DQB1
∗
03:01, DQB1
∗
03:02 and DQB1
∗
06:02/3 alleles
12
. The ca ie s o he geno ype DQB1
∗
02/DQB1
∗
03:02
o DQB1
∗
03:02/x geno ypes (he e x ≠DQB1
∗
02, DQB1
∗
03:01, DQB1
∗
06:02, o DQB1
∗
06:03 alleles) we e
ca ego ised in o he T1D isk g oup and ec ui ed o he DIPP ollow-up p og am.
De ec ion o be a-cell au oimmuni y
The pa icipan s wi h HLA-con e ed gene ic suscep ibili y we e moni o ed o he appea ance o T1D-
associa ed au oan ibodies: isle cell an ibodies (ICA), insulin au oan ibodies (IAA), isle an igen 2
au oan ibodies (IA-2A), and glu amic acid deca boxylase au oan ibodies (GADA). These au oan ibodies
we e measu ed in he Diabe es Resea ch Labo a o y (Uni e si y o Oulu) om he plasma samples aken
a each ollow-up isi
13
. ICA we e de ec ed wi h he use o indi ec immunofluo escence, whe eas he
o he h ee au oan ibodies we e quan ified wi h he use o specific adiobinding assays
14
. We used cu o
limi s o posi i i y o 2.5 Ju enile Diabe es Founda ion (JDF) uni s o ICA, 3.48 ela i e uni s (RU)
o IAA, 5.36 RU o GADA, and 0.43 RU o IA-2A. The disease sensi i i y and specifici y o he
assay o ICA we e 100% and 98%, espec i ely, in he ou h ound o he in e na ional wo kshops on
s anda disa ion o he ICA assay. Acco ding o he Diabe es Au oan ibody s anda disa ion P og am
12O subjec s wi hou
T1D o au o-an ibodies
(Co d blood)
6 mon hs
3 mon hs
Day 0
12 mon hs
24 mon hs
36 mon hs o abo e
18 mon hs
HLA-geno yping
DQB1 Sample collec ion(n=428)
Be ween1998-2012
De ec ion o β-cell au oimmuni y
(ICA, GADA, IA-2, IAA)
40 PT1D
40 P1Ab
40 CTRL
Sample
p epa a ion
UPLC-MS analysis
Figu e 1. An o e iew o he s udy design. In his s udy, newbo n in an s a he Uni e si y Hospi al o
Tampe e we e sc eened o inc eased gene ic isk o T1D. This s udy includes h ee s udy g oups: (1) 40
T1D (PT1D) p og esso s, (2) 40 who de eloped a leas a single Ab du ing he ollow-up bu did no p og ess o
T1D (P1Ab), and (3) 40 con ols (CTRL) who emained au oan ibody nega i e du ing he ollow-up. The h ee
s udy g oups we e ma ched by HLA-associa ed diabe es isk, gende and pe iod o bi h. Fo each child, up o
six longi udinal samples we e ob ained co esponding o he ages o 3, 6, 12, 18, 24 and 36 (o abo e) mon hs.
Abb e ia ions: isle cell an ibodies (ICA), insulin au oan ibodies (IAA), isle an igen 2 (IA-2) au oan ibodies,
and glu amic acid deca boxylase au oan ibodies (GADA).
PT1D (n =40) P1Ab (n =40) CTR(n =40)
Gende (gi ls, boys) (14, 26) (14, 26) (14, 26)
Age a ime o diagnosis (yea s) (mean ±SD) 4.75 ±2.94 NA NA
Age a ime o fi s se ocon e sion (yea s) (mean ±SD) 1.34 ±0.58 3.05 ±2.50 NA
Table 1. An h opome ic cha ac e is ics o s udy popula ion.
www.na u e.com/sda a/
SCIENTIFIC DATA |5:180250 |DOI: 10.1038/sda a.2018.250 3
(DASP) and Isle Au oan ibody s anda disa ion P og am (IASP) wo kshop esul s in 2010–2015, disease
sensi i i ies o he IAA, GADA and IA-2A adio binding assays we e 36–62%, 64–88% and 62–72%,
espec i ely. The co esponding disease specifici ies we e 94–98%, 94–99% and 93–100%, espec i ely.
The me ada a in he Me aboligh s con ains in o ma ion on he ypes o au oan ibodies de ec ed in he
longi udinal plasma samples (Da a Ci a ion 1).
Sample p epa a ion and UHPLC-MS analysis
The sol en s and eluen modifie s used in his wo k we e pu chased om Sigma-Ald ich (S einheim,
Ge many): HPLC-g ade chlo o o m, LC-MS g ade ace oni ile (ACN), isop opanol (IPA), wa e (H
2
O),
me hanol (MeOH), ammonium ace a e (NH
4
Ac), analy ical g ade o mic acid (HCOOH) and sodium
chlo ide (NaCl).
The ollowing in e nal s anda ds we e pu chased o quali y con ol (QC) and calib a ion pu poses:
1,2-dihep adecanoyl-sn-glyce o-3-phosphoe hanolamine (PE(17:0/17:0)), N-hep adecanoyl-D-e y h o-sphin-
gosylphospho ylcholine (SM(d18:1/17:0)), N-hep adecanoyl-D-e y h o-sphingosine (Ce (d18:1/17:0)),
1,2-dihep adecanoyl-sn-glyce o-3-phosphocholine (PC(17:0/17:0)), 1-hep adecanoyl-2-hyd oxy-sn-glyce o-
3-phosphocholine (LPC(17:0)) and 1-palmi oyl-d31-2-oleoyl-sn-glyce o-3-phosphocholine (PC(16:0/
d31/18:1)) om A an i Pola Lipids, Inc. (Alabas e , Alabama, USA), 1,2-Dimy is oyl-sn-glyce o-3-
phospho(choline-d
13
) (PC(14:0/d13)) om Sigma-Ald ich and T ipalmi in-1,1,1-13C3 (TG(16:0/16:0/16:0)-
13C3) and T ioc anoin-1,1,1-13C3 (TG(8:0/8:0/8:0)-13C3) om La odan (Solna, Sweden). S ock solu ions
(1.0 mg mL
−1
) o he in e nal s anda ds we e p epa ed by dissol ing he analy es in CHCl
3
:MeOH (2:1, / ).
The wo king s anda d solu ions we e p epa ed by u he dilu ing he s ock solu ions in CHCl
3
:MeOH (2:1,
/ ) o achie e concen a ions o 250 ng mL
−1
and 3.5 μgmL
−1
.
Calib a ion cu es using 1-hexadecyl-2-(9Z-oc adecenoyl)-sn-glyce o-3-phosphocholine (PC(16:0e/
18:1(9Z))), 1-(1Z-oc adecenyl)-2-(9Z-oc adecenoyl)-sn-glyce o-3-phosphocholine (PC(18:0p/18:1(9Z))),
1-oc adecanoyl-sn-glyce o-3-phosphocholine (LPC(18:0)), 1-(1Z-oc adecenyl)-2-docosahexaenoyl-sn-
glyce o-3-phosphocholine (PC(18:0p/22:6)) and 1-s ea oyl-2-linoleoyl-sn-glyce ol (DG(18:0/20:4)) om
A an i Pola Lipids Inc., 1-Palmi oyl-2-Hyd oxy-sn-Glyce o-3-Phospha idylcholine (LPC(16:0)) om
La odan, and 1,2,3-T ihep adecanoylglyce ol (TG(17:0/17:0/17:0)) and 3β-Hyd oxy-5-choles ene
3-linolea e (ChoE(18:2)) om Sigma-Ald ich we e p epa ed o he ollowing concen a ion le els: 100,
500, 1000, 1500, 2000 and 2500 ng mL
−1
(in CHCl
3
:MeOH, 2:1, / ) including 250 ng mL
−1
o each QC
s anda d.
A o al o 428 plasma samples we e ex ac ed in andomised o de using a modified e sion o he
Folch p ocedu e: 10 μL o 0.9% NaCl, 40 μL o CHCl
3
:MeOH (2:1, / ) and 80 μL o he 3.5 μgmL
−1
wo king s anda ds solu ion we e added o 10 μL o each plasma sample. The samples we e o ex mixed
and incuba ed on ice o 30 min a e which hey we e cen i uged (9400 × g, 3 min, 4 °C). F om he
lowe laye o each sample, 60 μL was ans e ed o a glass ial wi h an inse and 60 μL o CHCl
3
:MeOH
(2:1, / ) and added o each sample. The samples we e e- andomised and s o ed a 80 °C un il analysis
on he UHPLC-QTOF-MS sys em.
The UHPLC sys em used was a 1290 Infini y sys em om Agilen Technologies (San a Cla a,
Cali o nia, USA). The sys em was equipped wi h a mul isample (main ained a 10 °C) using 10% DCM
in MeOH and ACN:MeOH:IPA:H
2
O (1:1:1:1, / / / ) + 0.1% HCOOH as needle wash solu ions
a e each injec ion o 7.5 s each, a qua e na y sol en manage and a column he mos a (main ained a
50 °C). Sepa a ions we e pe o med on an ACQUITY UPLC
®
BEH C18 column (2.1 mm × 100 mm,
pa icle size 1.7 μm) by Wa e s (Mil o d, USA). The flow a e was 0.4 mL min
−1
and he injec ion olume
was 1 μL. H
2
O + 1% NH
4
Ac (1 M) + 0.1% HCOOH (A) and ACN:IPA (1:1, / ) + 1% NH
4
Ac + 0.1%
HCOOH (B) we e used as he mobile phases o he g adien elu ion. The g adien was as ollows: om 0
o 2 min. 35-80% B, om 2 o 7 min. 80–100% B and om 7 o 14 min 100% B. Each un was ollowed by
a 7 min e-equilib a ion pe iod unde he ini ial condi ions (35% B).
The mass spec ome e coupled o he UHPLC was a 6550 iFunnel QTOF-MS om Agilen
Technologies in e aced wi h a dual je s eam elec osp ay (dual ESI) ion sou ce. Ni ogen gene a ed by a
ni ogen gene a o (PEAK Scien ific, Sco land, UK) was used as he nebulising gas a a p essu e o 21 psi,
as he d ying gas a a flow a e o 14 L min
−1
(a 193 °C) and as he shea h gas a a flow a e o 11 L min
−1
(a 379 °C). Pu e ni ogen (6.0) om P axai (F ede icia, Denma k) was used as he collision gas. The
capilla y ol age and he nozzle ol age we e kep a 3643 V and 1500 V, espec i ely. The e e ence mass
solu ion including ions a m/z 121.0509 and 922.0098 was p epa ed acco ding o ins uc ions by Agilen
and was in oduced o he mass spec ome e h ough he o he nebulise in he dual ESI ion sou ce using
a sepa a e Agilen se ies 1290 isoc a ic pump a a cons an flow a e o 4 mL min
−1
(spli o 1:100 be o e
he nebulise ). The acquisi ion mass ange was m/z 100–1700 and he ins umen was un in ex ended
dynamic ange mode wi h an app oxima e esolu ion o 30,000 FWHM measu ed a m/z 1521.9715
(which is included in he une mix u e) du ing calib a ion o he ins umen . MassHun e B.06.01
so wa e (Agilen ) was used o da a acquisi ion.
Da a p e-p ocessing
The da a p e-p ocessing was pe o med using MZmine 2.18.2
15
. He e, we ha e adhe ed o he da a
p ocessing s eps as sugges ed by he me abolomics s anda ds ini ia i e
16
. The ypical p e-p ocessing
wo kflow includes aw file impo , fil e ing/smoo hing, de ec ion o peaks, peak lis de-iso oping,
www.na u e.com/sda a/
SCIENTIFIC DATA |5:180250 |DOI: 10.1038/sda a.2018.250 4
alignmen , gap filling, in eg a ion o peaks, no malisa ion and, finally, peak/ ea u e iden ifica ion (Fig. 2,
and Supplemen a y Table S1). The ollowing s eps we e applied in he p ocessing: (1) c op fil e ing wi h a
m/z ange o 350–1700 m/z and a e en ion ime (RT) ange o 2.5 o 21.0 min, (2) mass de ec ion wi h a
noise le el o 750, (3) ch oma og am builde wi h a minimum ime span o 0.08 min, minimum heigh o
2250 and a m/z ole ance o 0.006 m/z o 10.0 ppm, (4) Ch oma og am decon olu ion using he local
minimum sea ch algo i hm wi h a 70% ch oma og aphic h eshold, 0.05 min minimum RT ange, 5%
minimum ela i e heigh , 2250 minimum absolu e heigh , a minimum a ion o peak op/edge o 1 and a
peak du a ion ange o 0.08 o 5.0, (5) iso opic peak g oupe wi h a m/z ole ance o 5.0 ppm, RT
ole ance o 0.05 min, maximum cha ge o 2 and wi h he mos in ense iso ope se as he ep esen a i e
iso ope, (6) peak fil e wi h minimum 12 da a poin s, a FWHM be ween 0.0 and 0.2, ailing ac o
be ween 0.45 and 2.22 and asymme y ac o be ween 0.40 and 2.50, (7) peak lis ow fil e keeping only
peak wi h a minimum o 1 peak in a ow, (8) join aligne wi h a m/z ole ance o 0.006 o 10.0 ppm and a
weigh o o 2, a RT ole ance o 0.1 min and a weigh o 1 and wi h no equi emen o cha ge s a e o ID
and no compa ison o iso ope pa e n, (9) peak lis ow fil e wi h a minimum o 53 peak in a ow (=10%
o he samples), (10) duplica e peak fil e wi h a m/z ole ance o 0.006 m/z o 10.0 ppm and a RT
ole ance o 0.1 min, (11) gap filling using he same RT and m/z ange gap fille algo i hm wi h an m/z
ole ance o 0.006 m/z o 10.0 ppm, (12) peak fil e using he same pa ame e s as unde s ep 6, (13) peak
lis ow fil e wi h a minimum o 265 peaks in a ow (=50% o he samples), enabling de ec ion o 1084
ea u es (14) iden ifica ion o lipids using a cus om da abase sea ch (232 iden ified ou o 1084 ea u es)
wi h an m/z ole ance o 0.006 m/z o 10.0 ppm and a RT ole ance o 0.1 min, (15) no malisa ion using
lipid-class-specific in e nal s anda ds (PE (17:0/17:0), SM (d18:1/17:0), Ce (d18:1/17:0), LPC (17:0), TG
(16:0/16:0/16:0)-13C3 and PC (16:0/d30/18:1)) using in-house de eloped R-sc ip , and (16) da a
impu a ion o missing alues wi h he hal o he minimum alue o each lipid.
Code a ailabili y
All he p e-p ocessing analyses we e pe o med using he publicly-a ailable so wa e package MZmine
2.18.2 and wi h pa ame e s as desc ibed in Supplemen a y Table S1.
Da a eco ds
The aw da a files in .mzML o ma a e deposi ed in he Me aboLigh s eposi o y (Da a Ci a ion 1).
Addi ionally, he deposi ed da a con ains he p e-p ocessed lipidomic da a as ob ained om MZmine
Fil e ing
Peak de ec ion
Remo ing o iso opes
Iden i ica ion o agmen s, adduc s and peak complexes
Alignmen
Gap illing
No maliza ion Expo
Da a
analysis
Iden i ica ion
Da a clean
up
Using join Aligne
Using Linea no malize
SM(d18:1/17:0)
LPC(17:0)
PC(16:0/d30/18:1)
PE(17:0/17:0)
SM(d18:1/17:0)
TG(16:0/16:0/16:0)- C
m/z ange: 350 – 1700
RT ange: 2.5 – 21.0 min
50% minimum ea u es
To al: 1084 ea u es
232 known lipid species
Figu e 2. The ypical wo kflow o p ocessing mass spec ome y da a using MZmine 2. The wo kflow
includes aw file impo , fil e ing/smoo hing, de ec ion o peaks, peak lis de-iso oping, alignmen , gap filling,
and in eg a ion o peaks, no malisa ion, and, finally, peak/ ea u e iden ifica ion.
www.na u e.com/sda a/
SCIENTIFIC DATA |5:180250 |DOI: 10.1038/sda a.2018.250 5

da a p ocessing, including he iden ified lipids. The associa ed da a we e cap u ed using he ISA-c ea o
package a ailable om Me aboLigh s. This manusc ip desc ibes he samples, da a collec ion, p ocessing
s eps and o e all s udy design.
E hical app o al and in o med consen
The e hics and esea ch commi ees o bo h he pa icipa ing uni e si y and hospi al a Uni e si y o
Tampe e, Tampe e Finland, app o ed he s udy p o ocol. The s udy was conduc ed acco ding o he
guidelines in he Decla a ion o Helsinki. All amilies p o ided w i en in o med consen o pa icipa ion
in he s udy.
Technical alida ion
This s udy ollowed he bes p ac ices o analy ical me hodologies o he global p ofiling o lipids in
heplasmasampleasdesc ibedbyHyo ylainene al.
17,18
. Fo quali y moni o ing, QC samples (pooled
plasma samples), blank samples, pu e s anda ds (s anda ds in sol en ) we e un a egula in e als
a e e e y 4–16 s udy samples. To assu e da a quali y, ela i e s anda d de ia ions (%RSDs) we e
calcula ed o he e en ion imes and he peak a eas o each QC s anda d in he QC samples
0
0.5
1.0
1.5
2.0
2.5
3.0
3.5
4.0
4.5
5.0
050 100 150 200 250 300 350 400
0
2
4
6
8
10
12
14
16
18
PE(17:0/17:0)
SM(d18:1/17:0)
Ce (d18:1/17:0)
PC(17:0/17:0)
LPC(17:0)
(PC(14:0/d13)
(TG(16:0/16:0/16:0)-13C3)
(TG(8:0/8:0/8:0)-13C3)
PC(16:0/d31/18:1)
PC(16:0/d30/18:1)
Sample
050 100 150 200 250 300 350 400
a
b
Figu e 3. Quali y con ol o UHPLC-QTOFMS un e alua ed using peak a eas om nine in e nal
s anda d compounds. (a) Plo o peak a eas wi h 430 samples. (b) Plo o peak a eas o 428 samples, i.e. a e
emo al o 2 ou lie samples. He e, each colo ep esen s he peak a ea o he gi en in e nal s anda d PE
(17:0/17:0), SM(d18:1/17:0), Ce (d18:1/17:0), PC(17:0/17:0), LPC(17:0), PC(14:0/d13), TG(16:0/16:0/16:0)-
13C3, TG(8:0/8:0/8:0)-13C3 and PC(16:0/d31/18:1).
www.na u e.com/sda a/
SCIENTIFIC DATA |5:180250 |DOI: 10.1038/sda a.2018.250 6
h oughou he sample se . O e all, hese esul s show a ep oducible me hod. Fo example, pooled
plasma samples analysed h oughou he sample se show a e age %RSDs o he ou ba ches o
samples o less han 0.2 and 15.4% (no malised using PC(16:0/d30/18:1)) o he e en ion imes and
peak a eas, espec i ely. The %RSDs o he es o he QC samples we e in he same ange as o he
pooled plasma samples.
To assess expe imen al e o , nine in e nal s anda ds we e spiked in o he plasma samples du ing sample
p epa a ion. The plo o peak a eas om nine QC s anda ds (Fig. 3a) clea ly shows ha wo samples had
eccen ic in e nal s anda d p ofiles. This was aced o an e o du ing sample p epa a ion, whe ein he QC
s anda d was added wo hese wo samples wice. The e o e, hese wo ou lie samples we e no included in
downs eam analysis (and no included in he da a desc ip o ). The ep esen a i e quali y plo s a e
emo al o abe an samples (Fig. 3b) demons a es ha he peak a eas o he QC s anda ds ha e e y low
a iabili y be ween he measu ed samples. This indica es ha use o p ope in e nal s anda ds and ca e ul
ollow-up o he esponse o hese in e nal s anda ds c i ically educes he chances o unexpec ed a ia ion,
which may a ise in LC-MS analysis. Based on his, we conclude ha he samples in his da a desc ip o we e
success ully analysed ollowing gene ally accep ed analy ical quali y guidelines.
Du ing LC-MS analyses, i is also highly impo an o keep ack o blank samples inse ed in o he
sample sequence. This is done o de ec possible con amina ion/in e e ence, which a ec he ac ual
sample da a. We analysed 293 blank o sol en blank samples h oughou his sample se . Sol en
blank samples (consis ing o a mix u e o CHCl
3
: MeOH, 2:1, / ) we e analysed a e e e y 4
h
sample and
blank samples ( he same sol en mix u e, which unde wen he no mal sample p epa a ion p ocedu e
excep o he addi ion o in e nal s anda ds) a e e e y 16
h
sample. Addi ionally, a se o QC samples (186
samples) we e analysed a he same equency as he blank samples (i.e. a e e e y 16
h
sample). This se o
QC samples consis ed o : (1) ex ac ed and non-ex ac ed s anda d samples ( he chosen QC s anda ds in
sol en ), (2) a pooled plasma sample o igina ing om a la ge plasma pool om ac ual diabe es
pa ien s o ollowing up on long- e m me hod pe o mance and o acili a e ba ch co ec ion i needed, and
(3) a S anda d Re e ence Ma e ial 1950 sample om he Na ional Ins i u e o S anda ds and Technology
(NIST) ha se e as communi y-wide benchma ks o in a- and in e -labo a o y QC and me hod
alida ion
19
.
Nex , o quan i a i e analysis, calib a ion cu es o each analy e mus be gene a ed. Fo global
lipidomics analyses his is, howe e , logis ically and p ac ically impossible due o he high numbe o
analy es p esen in he biological sample ma ix. To o e come his limi a ion, lipid-class-specific
compounds we e used o c ea ing a semi-quan i a i e me hod including six di e en concen a ions
o he lipid class specific s anda ds PC (16:0e/18:1(9Z)), TG (17:0/17:0/17:0), LPC (18:0), DG
(18:0/20:4) and CE (18:2). The accep ance o LC-MS da a usually elies on QC samples being
Concen a ion ng/ml
1000
500
1500
2000
2500
0 5 10 15 20
No malized Ra ios
LPC (16:0) [R : 0.998]
02 46 81012
1000
500
1500
2000
2500
Concen a ion ng/ml
TG (17:0/17:0/17:0) [R : 0.988]
No malized Ra ios
Figu e 4. Rep esen a i e global linea models. Six di e en concen a ions o lipid class-specific s anda ds
TG (17:0/17:0/17:0), and LPC (16:0) we e used o gene a e he calib a ion plo .
www.na u e.com/sda a/
SCIENTIFIC DATA |5:180250 |DOI: 10.1038/sda a.2018.250 7
success ully quan ified wi hin p e-defined limi s. Achie emen o his s anda d is e alua ed
s a is ically (global linea models) based on he alue o he coe ficien o de e mina ion (R
2
).
Figu e 4 shows coe ficien s o de e mina ion highe han 0.9 o TG (17:0/17:0/17:0) and LPC(16:0) a
six di e en concen a ions using a global linea model wi h a 1/x weigh ing. The o he s anda ds
showsimila alues(i.e.R
2
>0.9). This s a is ical analysis consequen ly ensu ed he eliabili y o
esul s p esen ed in his da a desc ip o .
Fu he mo e, he assu ances p o ided by he QC me hods he ein de i e om analy ical
ep oducibili y, wi hin o be ween ba ches o he LC-MS da a. These in e - and in a-ba ch a ia ions
can be assessed and co ec ed using he QC samples. The p incipal componen analysis (PCA) sco e plo
(Supplemen a y Figu e S1) highligh s he a o emen ioned ep oducibili y ac oss all ba ches ob ained
om he p e-p ocessed da ase in his da a desc ip o . F om his sco e plo , i is appa en ha he e a e
no pa e ns o any clus e ing ha co ela ed wi h he ba ch un du ing LC-MS. Toge he , he da ase in
his da a desc ip o eflec s obus ness o lipid measu emen o e ime, which, in u n, indica es ha da a
analysis pe o med using his da ase will no be a ec ed by sys ema ic e o a ising due o ins umen al
ac o s.
Usage no es
The p e-p ocessed da a a e a ailable om Me aboLigh s (Da a Ci a ion 1). The de ail pa ame e s o
da a p e-p ocessing a e a ailable in supplemen a y ma e ial (Supplemen a y Table S1). The p e-
p ocessed da a we e no malised using he nine di e en class-specific in e nal s anda ds. The class-
specific we e based on PE (17:0/17:0), SM (d18:1/17:0), Ce (d18:1/17:0), LPC (17:0), TG
(16:0/16:0/16:0)-13C3 and PC (16:0/d30/18:1), while o he s we e based on e en ion ime; LPC
(17:0) o o=6.0 min elu ions, PC(16:0/d30/18:1) o >6.0, o9.0 min elu ions and TG(16:0/16:0-
/16:0)-13C3 o >=9.0 min elu ions. In addi ion, he ze o alues in he lipidomics da ase we e
impu ed wi h hal o ha ow’s minimum alue o each lipid ( he code o his impu a ion we e
w i en in-house using R, and a e a ailable on eques ). Due o he non-no mal dis ibu ion o da a, i
is ecommended ha he da a a e log- ans o med be o e analysis (log- ans o ma ion pe o med in
MZmine 2.18.2). O he equi ed guidelines o da a-p ep ocessing a e a ailable om he MZmine
websi e (h p://mzmine.gi hub.io/documen a ion.h ml) and au ho s can be con ac ed o u he help
i equi ed.
Re e ences
1. Achenbach, P., Boni acio, E., Koczwa a, K. & Ziegle , A.-G. Na u al his o y o ype 1 diabe es. Diabe es 54, S25–31 (2005).
2. Pa e son, C. C., Dahlquis , G. G., Gyü üs, E., G een, A. & Sol ész, G. Incidence ends o childhood ype 1 diabe es in Eu ope
du ing 1989–2003 and p edic ed new cases 2005–20: a mul icen e p ospec i e egis a ion s udy. Lance 373, 2027–2033 (2009).
3. Kimpimaki, T. e al. Na u al his o y o be a-cell au oimmuni y in young child en wi h inc eased gene ic suscep ibili y o ype 1
diabe es ec ui ed om he gene al popula ion. J. Clin. Endoc inol. Me ab. 87, 4572–4579 (2002).
4. Ziegle , A. G. e al. Se ocon e sion o mul iple isle au oan ibodies and isk o p og ession o diabe es in child en. JAMA 309,
2473–2479 (2013).
5. La To e, D. e al. Dec eased Co d-Blood Phospholipids in Young Age–a –Onse Type 1 Diabe es. Diabe es 62, 3951–3956 (2013).
6. O esic, M. e al. Co d se um lipidome in p edic ion o isle au oimmuni y and ype 1 diabe es. Diabe es 62, 3268–3274 (2013).
7. O esic, M. e al. Dys egula ion o lipid and amino acid me abolism p ecedes isle au oimmuni y in child en who la e p og ess o
ype 1 diabe es. J. Exp. Med. 205, 2975–2984 (2008).
8. Lamichhane, S. e al. Dynamics o Plasma Lipidome in P og ession o Isle Au oimmuni y and Type 1 Diabe es –Type 1 Diabe es
P edic ion and P e en ion S udy (DIPP). Sci. Rep. 8, 10635 (2018).
9. Kupila, A. e al. Feasibili y o gene ic and immunological p edic ion o ype I diabe es in a popula ion-based bi h coho .
Diabe ologia 44, 290–297 (2001).
10. Halle , M. J. & Scha z, D. A. The DIPP p ojec : 20 yea s o disco e y in ype 1 diabe es. Pedia . Diabe es 17(Suppl 22):
5–7 (2016).
11. Nejen se , S. e al. Popula ion-based gene ic sc eening o he es ima ion o Type 1 diabe es melli us isk in Finland: selec i e
geno yping o ma ke s in he HLA-DQB1, HLA-DQA1 and HLA-DRB1 loci. Diabe . Med 16, 985–992 (1999).
12. Ilonen, J. e al. Rapid HLA-DQB1 geno yping o ou alleles in he assessmen o isk o IDDM in he Finnish popula ion. The
Childhood Diabe es in Finland (DiMe) S udy G oup. Diabe es Ca e 19, 795–800 (1996).
13. Siljande , H. T. e al. P edic i e cha ac e is ics o diabe es-associa ed au oan ibodies among child en wi h HLA-con e ed disease
suscep ibili y in he gene al popula ion. Diabe es 58, 2835–2842 (2009).
14. Knip, M. e al. Die a y in e en ion in in ancy and la e signs o be a-cell au oimmuni y. N. Engl. J. Med. 363, 1900–1908 (2010).
15. Pluskal, T., Cas illo, S., Villa -B iones, A. & O esic, M. MZmine 2: modula amewo k o p ocessing, isualizing, and analyzing
mass spec ome y-based molecula p ofile da a. BMC Bioin o ma ics 11, 395 (2010).
16. Goodac e, R. e al. P oposed minimum epo ing s anda ds o da a analysis in me abolomics. Me abolomics 3, 231–241 (2007).
17. Hyo ylainen, T., Ahonen, L., Poho, P. & O esic, M. Lipidomics in biomedical esea ch-p ac ical conside a ions. Biochim. Biophys.
Ac a 1862, 800–803 (2017).
18. Hyo ylainen, T. & O esic, M. Bioanaly ical echniques in non a ge ed clinical lipidomics. Bioanalysis 8, 351–364 (2016).
19. Bowden, J. A. e al. Ha monizing lipidomics: NIST in e labo a o y compa ison exe cise o lipidomics using SRM 1950-
Me aboli es in F ozen Human Plasma. J. Lipid Res. 58, 2275–2288 (2017).
Da a Ci a ions
1. Lamichhane, S., Knip, M. & O esic, M. Me aboLigh s MTBLS620 (2018).
Acknowledgemen s
This wo k was suppo ed by he JDRF g an s 4-1998-274, 4-1999-731 4-2001-435 and special esea ch
unds o Oulu, Tampe e and Tu ku Uni e si y Hospi als in Finland. This wo k was suppo ed by he
www.na u e.com/sda a/
SCIENTIFIC DATA |5:180250 |DOI: 10.1038/sda a.2018.250 8
Ju enile Diabe es Resea ch Founda ion (2-SRA-2014-159-Q-R o M.O.) and he Academy o Finland
(Cen e o Excellence in Molecula Sys ems Immunology and Physiology Resea ch –SyMMyS, Decision
No. 250114, o M.O. and M.K.). We gi e ou hanks o Olli Simell o his con ibu ion o he DIPP s udy,
Ane e Un e mann o excellen echnical suppo in me abolomics analysis and Aidan McGlinchey o
his con ibu ion in p oo eading he manusc ip .
Au ho Con ibu ions
M.O. and M.K. designed and supe ised he s udy. L.A. and T.S.D. pe o med me abolomics analysis,
which was supe ised by T.H.; H.H., J.I., J.T. and R.V. con ibu ed o he design o he clinical s udy. S.L.
and M.O. w o e he manusc ip . All au ho s c i ically e iewed and app o ed he final manusc ip .
Addi ional In o ma ion
Supplemen a y in o ma ion accompanies his pape a h p://www.na u e.com/sda a
Compe ing in e es s: The au ho s decla e no compe ing in e es s.
How o ci e his a icle: Lamichhane, S. e al. A longi udinal plasma lipidomics da ase om child en
who de eloped isle au oimmuni y and ype 1 diabe es. Sci. Da a. 5:180250 doi: 10.1038/sda a.2018.250
(2018).
Publishe ’s no e: Sp inge Na u e emains neu al wi h ega d o ju isdic ional claims in published maps
and ins i u ional a filia ions.
Open Access This a icle is licensed unde a C ea i e Commons A ibu ion 4.0 In e na-
ional License, which pe mi s use, sha ing, adap a ion, dis ibu ion and ep oduc ion in any
medium o o ma , as long as you gi e app op ia e c edi o he o iginal au ho (s) and he sou ce, p o ide a
link o he C ea i e Commons license, and indica e i changes we e made. The images o o he hi d pa y
ma e ial in his a icle a e included in he a icle’s C ea i e Commons license, unless indica ed o he wise in
a c edi line o he ma e ial. I ma e ial is no included in he a icle’s C ea i e Commons license and you
in ended use is no pe mi ed by s a u o y egula ion o exceeds he pe mi ed use, you will need o ob ain
pe mission di ec ly om he copy igh holde . To iew a copy o his license, isi h p://c ea i ecommons.
o g/licenses/by/4.0/
The C ea i e Commons Public Domain Dedica ion wai e h p://c ea i ecommons.o g/publicdomain/
ze o/1.0/ applies o he me ada a files made a ailable in his a icle.
© The Au ho (s) 2018
www.na u e.com/sda a/
SCIENTIFIC DATA |5:180250 |DOI: 10.1038/sda a.2018.250 9