RESEARCH Open Access
© The Au ho (s) 2025. Open Access This a icle is licensed unde a C ea i e Commons A ibu ion-NonComme cial-NoDe i a i es 4.0
In e na ional License, which pe mi s any non-comme cial use, sha ing, 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 licence, and indica e i you modi ied he
licensed ma e ial. You do no ha e pe mission unde his licence o sha e adap ed ma e ial de i ed om his a icle o pa s o i . 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 licence, 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 licence 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 licence, isi h p : / / c e a i
e c o m m o n s . o g / l i c e n s e s / b y - n c - n d / 4 . 0 /.
Pa yi e al. BMC Plan Biology (2025) 25:945
h ps://doi.o g/10.1186/s12870-025-06951-7 BMC Plan Biology
*Co espondence:
And ás Pa yi
[email p o ec ed]
Full lis o au ho in o ma ion is a ailable a he end o he a icle
Abs ac
Backg ound Whi e lupin (Lupinus albus, L.) is a g ain legume cul i a ed o i s high ag onomic and nu i ional
po en ial, despi e he accumula ion o bi e and po en ially ha m ul o heal h quinolizidine alkaloids (QA) in he g ain.
Mode n swee (i.e. low alkaloid con en ) a ie ies exis , which a e exploi ing di e en ecessi e mu a ions esponsible
o he desi ed low QA chemo ype. The mos widely used QA- educing de e minan , paupe , has been ecen ly
iden i ied enabling ma ke -assis ed selec ion, bu uns able QA con en ac oss g owing seasons and en i onmen s
emain a challenge in whi e lupin cul i a ion.
Resul s Th ough Bulked Seg egan Analysis o an F2 popula ion, seg ega ing o wo di e en low QA con e ing
loci, we iden i ied a no el QTL spanning a 1 Mbp egion on ch omosome 5, a no el sou ce o swee ness apa om
paupe . We p esen ma ke - ai associa ions o he new locus agging low QA con en in whi e lupin wi hin his
QTL, desc ibed in F2 gene a ion and alida ed in F3. Toge he wi h geno yping o swee ness agging paupe swee
indi iduals, we iden i ied, in F3, 50 s acked allele ecombinan s whe e he low QA chemo ype is u he educed. These
indi iduals exhibi an excep ionally low o al alkaloid con en (22.8 ± 10.4 ppm), e en when compa ed o geno ypes
known o ca y he allele associa ed o he mos d as ic educ ion in QAs, paupe (171.7 ± 18.5 ppm).
Conclusion The disco e y o his no el locus and he de elopmen o associa ed PACE ma ke s, po en ially
applicable o ma ke -assis ed selec ion oge he wi h paupe (especially a e a he alida ion in a la ge panel o
accessions), can enhance he imp o emen o whi e lupin h ough he de elopmen o new a ie ies wi h e y low
Iden i ica ion o a new QTL associa ed
o educed quinolizidine alkaloid
con en in whi e lupin (Lupinus albus, L.)
and de elopmen o ul a-low alkaloid
ecombinan s by s acking wi h he paupe
allele
And ásPa yi1,2*, Mi iamKamp3, Ch is ineA ncken1, ElisaBiazzi4, MichałKsiążkiewicz2, Monika M.Messme 1,
MichaelSchneide 1, AldoTa a4 and Ma ia e esaLazza o1
Page 2 o 14Pa yi e al. BMC Plan Biology (2025) 25:945
Backg ound
Whi e lupin (Lupinus albus, L.) is a g ain legume o Med-
i e anean o igin p oduced o i s ag onomic alue and
highly nu i ious seeds. F om he nu i ional pe spec i e,
whi e lupin g ain is app ecia ed o eed and ood o i s
high p o ein (33–47% depending on geno ype and loca-
ion) and excellen ib e con en [1]. I s desi able a y acid
composi ion a io o omega-3 o omega-6, non-s a ch
ca bohyd a e-, oligosaccha ide- and an ioxidan -con en
makes he consump ion o whi e lupin seed a g ea con-
ibu ion o a heal hy die [2]. Con e sely, whi e lupin
seed con ains quinolizidine alkaloids (QAs), oxic (p i-
ma ily due o hei abili y o in e e e wi h he ne ous
sys em) and bi e compounds which se a majo chal-
lenge o ood p oduce s and p ocesso s. In he closely-
ela ed blue lupin (Lupinus angus i olius, L.), he majo
si e o p oduc ion o QAs is he ae ial epide mis, la e
anspo ed o seeds du ing ui de elopmen [3]. Seeds
o wild o ms o whi e lupin con ain up o 127,000 ppm
(mg/kg d y weigh ) QAs which is educed in bi e whi e
lupin cul i a s o 2,000–4,000 ppm [4]. His o ically, he
g ains we e soaked in unning wa e o emo e he oxic
alkaloids be o e consump ion. Imp o ed me hods exis
o debi e ing, commonly using wa e as a sol en , how-
e e hese p ocesses emain wa e -use in ensi e, labo i-
ous and p oducing oxic was e [5], cons aining he use
and economic iabili y o whi e lupin in he alue chain,
o eed and, especially, ood.
Lupins con ain a mix u e o di e en alkaloids, no
all o which a e su icien ly known o hei oxici y.
Recen ly, Sch eibe e al. [6] e alua ed he oxici y o
i e common lupin alkaloids (spa eine, lupinine, lupa-
nine, 13-hyd oxylupanine, angus i oline) epo ing ha
hese QAs indi idually and as a mix u e a e nei he
cy o-, no geno oxic in mammalian cells. Conce ning
hei neu o oxici y, he e ec o spa eine, p esumably
and s able alkaloid con en based on s acked allele ecombinan s. This can help o inc ease he cul i a ion o his
use ul ye unde u ilised c op and i s use o human nu i ion.
Keywo ds Whi e lupin, Bulked Seg egan analysis (BSA), Quan i a i e ai locus (QTL), Quinolizidine alkaloids (QA),
PCR allelic compe i i e ex ension (PACE), Plan b eeding
G aphical abs ac
Page 3 o 14Pa yi e al. BMC Plan Biology (2025) 25:945
he mos oxic lupin alkaloid, has been s udied bes .
The a ailable isk assessmen is he e o e based on
he knowledge o spa eine o all alkaloids p esen in
lupins. Sa e y con ol is based on he o al quinolizidine
alkaloid con en [7]. In Swi ze land, o example, he e
a e no legal limi s o alkaloids in ood o eed. The
p inciple o sel - egula ion applies, which means ha
companies mus ensu e ha only sa e ood is placed on
he ma ke [8]. The Ge man Fede al Ins i u e o Risk
Assessmen (B R) has es ablished in 2017 he ollowing
guideline alues o p o ec human and animal heal h:
<200 ppm and < 500 ppm (d y weigh ) espec i ely o
ood and eed, ela i e o he inal p oduc in ended o
consump ion [9]. Addi ionally, he impo and expo
s anda ds o he indus y o ganisa ion Pulse Aus a-
lia (Aus alia being he wo ld’s la ges lupin p oduce )
se in 2001 a maximum alkaloid con en o 200 ppm
[7]. The Eu opean Commission has ecen ly ecom-
mended moni o ing o he ollowing quinolizidine
alkaloids in lupins and lupin-de i ed ood: albine, ana-
gy in, angus i oline, lupanine, isolupanine, mul i lo ine,
13α-hyd oxylupanine, lupinine and spa eine [10].
Mode n swee (i.e. bi e ness lacking) a ie ies exis
wi h he alkaloid con en anging be ween 100 and
800 ppm o seed d y weigh [11]. Low QA cul i a s
we e a majo achie emen o he o mal b eeding o
whi e lupin, s a ed in he ea ly 20 h cen u y wi h he
ac i i y o Reinhold on Sengbusch [12]. He selec ed
agains common wild ype ai s, amongs many o he s
o high QA con en , desc ibing he d as ic dec ease ( o
200–500 ppm) in QA con en due o he ecessi e allele
paupe , p esen in many mode n whi e lupin cul i a s
[13]. Apa om paupe , o he loci exis (e.g. exiguus,
nu icius, mi is, educ us) ep esen ing possible eces-
si e mu a ions o low QA con en , desc ibed o educe
QA con en o ≤ 1000 ppm [11]. These mu a ions we e
desc ibed based on c ossings be ween di e en whi e
lupin indi iduals and obse ing he seg ega ion pa e n
in he chemo ype, hus no much is known abou he
genes in ol ed o ansc ip omic egula ion o low QA
con en in whi e lupins, excep o paupe . Recen ly, he
mu a ion causing he paupe - ype swee ness in whi e
lupin was iden i ied [14]. Repo edly he p esence o a
single SNP (Lalb_Ch 18_12359687) con e s a g ea ly
educed o al QA con en and a dis inc QA p o ile. This
mu a ion is loca ed in he coding sequence o a gene
encoding an ace yl ans e ase, and is esponsible o
changes in he ea ly biosyn he ic s eps o QAs, esul -
ing in he low QA chemo ype. Howe e , pheno ypically
swee a ian s exis , which do no ca y he paupe
mu a ion, con i ming non-paupe swee ness ypes, ea -
lie desc ibed in he 20 h cen u y [11]. One example
was he cul i a ‘Die a’, in which he gene ic de e mi-
nan causing low QA le els is unknown [14].
Despi e he magni ude o he e ec o he gene associ-
a ed wi h paupe , g ains ha es ed om epo edly swee
paupe accessions ha e been obse ed o exceed he
ad ised h eshold o consump ion (FiBL, unpublished
da a). Rela i ely high and uns able QA con en in he
ha es ed g ain emains a h ea o whi e lupin p oduc-
e s and p ocesse s, c ea ing a demand o e en lowe QA
a ie ies wi h s able swee ness ac oss g owing seasons
ega dless o he en i onmen al con ex . This could be
achie ed h ough gene s acking (py amiding)– combin-
ing independen loci con e ing low QA con en in whi e
lupin, which could be u he aided wi h he a ailabili y
o molecula ma ke s agging hese loci.
The p esen s udy aimed o map an unknown sou ce o
swee ness in whi e lupin by Bulked Seg egan Analysis
(BSA) in an F2 popula ion esul ing om he c ossing o
he paupe swee cul i a ‘F ieda’ and he cul i a ‘Die a’
ca ying he unknown swee ness de e minan . Subse-
quen ly, h ough PACE ma ke de elopmen and alida-
ion in a di e si y panel, we aimed o e i y he e ec o
he indi idual loci con e ing he low QA le el pheno ype
and o assess hei combined e ec on QA concen a ion
and p o ile in he F2 and he de i ed F3 popula ion.
Me hods
Plan ma e ial
C ossing and F2 popula ion
Two comme cial cul i a s, ‘F ieda’ (♂, paupe swee -
ness) and ‘Die a’ (♀, unknown swee ness), we e c ossed
in 2021. Seeds we e ob ained om Delley Samen und
P lanzen AG (Delley-Po alban, Swi ze land) o ‘F ieda’
and om Soya UK L d. (Sou hamp on, Uni ed Kingdom)
o ‘Die a’. The i e ha es ed seeds om he c ossing
we e p opaga ed in 2022 unde p o ec i e ne s o p e en
c oss-pollina ion and seeds o he F1 plan s we e ha -
es ed in a bulk.
The F2 plan s (own b eeding ma e ial) we e g own in
2023 in he FiBL b eeding nu se y loca ed in Leibs ad ,
Swi ze land (47°35’26.4"N 8°10’56.0"E), o ganized in
a 10- ow plo wi h 14 F1:2 seeds sown pe ow, g own
unde p o ec i e ne s. Young lea issue (ca. 100mg) was
collec ed and ozen om he 112 su i ing indi iduals
o he F2 popula ion (Supplemen a y Table S1).
F3 popula ion
Thi een F2 plan s, all he e ozygous a he paupe locus
we e selec ed o he F3 popula ion (own b eeding ma e-
ial) and hei seeds we e g own in 2024 on a ield in
Full-Reuen hal, Swi ze land (47°35’59.7”N 8°12’13.6”E),
sown and p ocessed in an iden ical manne as in p e ious
yea s. In o al, 329 indi iduals ep esen ed he F3 popula-
ion (Supplemen a y Table S3). In addi ion, 10 indi idu-
als each o ‘F ieda’ and ‘Die a’ we e g own in he same
ield and ha es ed as single plan s.
Page 4 o 14Pa yi e al. BMC Plan Biology (2025) 25:945
Valida ion panel
A di e se panel o 34 indi iduals was c ea ed o he
molecula ma ke alida ion, consis ing o 1 o 7 ep e-
sen a i es o 1 b eeding line, 16 cul i a s, 3 land aces
and 1 wild ype in addi ion o he 39 accessions o he
whi e lupin pangenome [15] (Supplemen a y Table S5).
Pheno yping
The alkaloid con en o F2 and F3 was es ima ed i s on
he ield using D agendo pape es (Supplemen a y
Table S1, Supplemen a y Table S3), a il e pape soaked
in iodine/po assium iodide solu ion (D agendo solu-
ion), hen d ied a 40°C [16]. Indi idual QA con en was
es ima ed by ans e ing plan sap om he pe ioles on o
he D agendo pape du ing onse o lowe ing (BBCH
63–65). A isible eac ion o he D agendo pape
u ning a eddish-b own colou was desc ibed o occu
o QA con en s o ≥ 500 ppm [17]. The colou eac ion
was sco ed om 1 o 4, 1 meaning a pu ely whi e s ain,
he lowes alkaloid con en . A sco e o 2 was assigned o
s ains, which we e mos ly whi e wi h he ed enci cling,
sco e 3 desc ibed ed enci cled pinkish s ains wi h a ligh
cen e and las ly, a sco e o 4 was gi en o samples wi h
an in ense ed s ain wi hou discolou a ion in he cen e
(Fig.1).
Addi ionally, bi e ness in he F2 popula ion was phe-
no yped by sco ing he as e o one lowe pe plan in
a ange om 1 (absence o bi e ness) o 9 (ex emely
bi e ). Fo a mo e accu a e pheno ypic sco e in he F2,
we combined he D agendo pheno ypic sco e mul i-
plied by wo o ha o he bi e ness as ing sco e. In his
combined pheno ypic sco e, he pheno ypic alue anged
om 3 o 15, 15 being he mos bi e .
Chemo yping o F3 indi idual seeds
App oxima ely 1g o seeds (3–5 seeds) o a cu a ed panel
o F3 (Supplemen a y Table S6) and pa en al indi iduals
we e milled o analy ical QA quan i ica ion. QA ex ac-
ion and cha ac e iza ion we e pe o med acco ding o
modi ied e sions o he p o ocol desc ibed in Wink [18],
Res a e al. [19] and Boschin e al. [20]. B ie ly, 100mg
o seed lou was suspended in 1.2 mL o 0.1N HCl,
wi h spa eine (CAS 90-39-1; Ex asyn hese, F ance)
added as in e nal s anda d in an app op ia e concen a-
ion, and s i ed a oom empe a u e o e nigh . The
mix u e was cen i uged a 8000g o 45min a 4°C,
he supe na an was collec ed, and he solid was washed
wice wi h 0.8 mL o 0.1N HCl. The ga he ed ex ac s
we e alkalinized wi h 5% NH4OH o pH 10 − 11 and
hen applied on o an Ex elu NT 3 column (Supelco,
Me ck, Da ms ad , Ge many). The alkaloids we e elu ed
wi h CH2Cl2 (3 × 5 mL), and he sol en was e apo-
a ed unde acuum. The esidue was esuspended in an
app op ia e olume o dichlo ome hane and analyzed
Fig. 1 Examples o he colou - eac ion o D agendo pape acco ding o he p ede ined sco ing scale om 1 ( e y swee ) o 4 ( e y bi e ). 1 - pu ely
whi e s ain; 2 - mos ly whi e wi h he ed enci cling; 3 - ed enci cled pinkish s ains wi h a ligh cen e 4 - in ense ed s ain wi hou discolou a ion in he
cen e
Page 5 o 14Pa yi e al. BMC Plan Biology (2025) 25:945
by gas ch oma og aphy. Each sample was independen ly
ex ac ed, hen analyzed in duplica e.
To al and indi idual QA con en was de e mined
quan i a i ely using he in e nal s anda d me hods by
gas ch oma og aphy/Flame Ioniza ion De ec o (Pe kin-
Elme Cla us 500 GC-FID), equipped wi h a capilla y
column Eli e-5 MS (DB-5, 30m × 0.32mm × 0.25μm;
Pe kin-Elme ; Milan, I aly). The o en empe a u e amp
was held a 90°C o 2min, inc eased o 300°C a 7°C/
min and held a 300°C o 10min. Helium was used as
he ca ie gas, and he low a e was se a 2 mL/min.
The empe a u e o he injec o was se a 300°C, and he
injec ion olume was 1 µL; he empe a u es o he FID
was se a 320°C. The absence o spa eine in he sam-
ples was demons a ed by pe o ming GC/FID analyses
wi hou an in e nal s anda d (choosing one sample pe
ecombinan ).
The esponse ac o o GC/FID was calcula ed using
he a io be ween he esponse o he in e nal s anda d
(spa eine) and he esponse o he analy e s anda d lupa-
nine. The eg ession coe icien be ween he analy e con-
cen a ion and de ec o esponse was R2 = 0.991.
Quali a i e iden i ica ion o QAs was pe o med by
gas ch oma og aphy/mass spec ome y; GC/MS analy-
ses we e ca ied ou using a Pe kin Elme Cla us 500 GC
equipped wi h a Cla us 500 mass spec ome e using he
same capilla y column and ch oma og aphic condi ions
as o he GC/FID analyses (abo e men ioned). Mass
spec a we e acqui ed o e a ange o 40–400 a omic
mass uni s (amu) a 1 scan/sec wi h ionizing elec on
ene gy o 70eV and ion sou ce a 230°C. The ans e
line was se a 300°C, while he ca ie gas was helium
a 1.0 mL/min. The QAs we e iden i ied by de e mina-
ion o hei elu ion imes published mass spec a [18–
20], as well as a peak-ma ching lib a y sea ch [21]. The
s anda d GC condi ion desc ibed abo e allowed a p ac i-
cal measu able sensi i i y o 10 ng/µl pe injec ion, i.e. a
de ec ion limi o 1µg o lupanine g−1 o lupin seed lou
(1 ppm).
Pooling o bulked seg egan analysis
Two bulks consis ing o DNA (isola ed wi h a modi ied
CTAB me hod) om eigh plan s o he swee and bi e
pheno ypic ex emes o he ‘Die a’ x ‘F ieda’ F2 expe i-
men al popula ion we e cu a ed o sequencing. Homo-
zygous indi iduals o he paupe locus we e excluded
om his subse o disc imina e swee indi iduals whe e
he low QA con en is con e ed by paupe . Thus, he
F2 bulks consis ed o indi iduals which we e he e ozy-
gous o wild ype a he paupe locus, sugges ing ha he
low QA pheno ype is con e ed by a de e minan o he
han paupe . The combined pheno ypic sco e [3–15] o
he swee bulk was be ween 7 and 8.5, while ha o he
bi e one was 13–15. Addi ionally, bulks o he pa en al
cul i a s ( om seedlings g own o his pu pose om
he same seed sou ce used o he c ossing) consis ing o
eigh plan s each we e sen o sequencing.
Sequencing-da a analysis
Whole genome sequencing was pe o med a No ogene
UK (Camb idge, UK) wi h a No aSeq X Plus Se ies,
sequencing dep h o 15x. Pai ed-end 150bp long eads
we e gene a ed. The aw sequence da a was p ocessed
and analysed in a modi ied e sion o he p ocedu e
published by Schneide e al. 2022 [22]. Quali y checks
we e pe o med wi h he as qc ool [23]. Raw eads we e
aligned o he e e ence genome using bwa mem [24].
Subsequen ly, duplica ed eads we e emo ed using he
ma kdub- unc ion o sambamba [25] and so ed by hei
mapped posi ion. Reads ha had been mapped o mul i-
ple loca ions (e.g. seconda y alignmen s) we e disca ded.
The a ian calling was conduc ed using bc ools only on
SNPs ha had been p e iously epo ed [15] o imp o e
he con idence despi e he low sample numbe and ead
dep h. Fu he mo e, eads wi h a quali y h eshold below
25 o he whole ead, and 30 pe SNV base call we e
omi ed, as well as SNPs wi h ead dep hs below 8. The
allelic dep h (AD) was ex ac ed and he e ozygous SNPs
in he pa en al bulks we e omi ed om u he analysis.
The emaining SNPs we e assigned o anno a ed genes
based on hei posi ion in he gene o in close p oxim-
i y o i , as desc ibed by Schneide e al. 2022 [22]. Fo
he esul ing 23,396 gene-ancho ed haplo ypes, he allele
equencies o ‘Die a’ and ‘F ieda’ we e calcula ed o he
swee and o he bi e bulk. The equencies we e agg e-
ga ed wi h he ead dep h o coun alues, as equi ed
o BayPass [26], which is a Bayesian app oach o iden-
i y i.e. selec i e sweeps be ween he swee and bi e
popula ions. BayPass was conduc ed using he e ec i e
popula ion size o 200 o e lec he size o he ‘Die a’ ×
‘F ieda’ F2 popula ion. F om he p- alues o each haplo-
ype and hei mapped posi ions (Supplemen a y Table
S8), Manha an plo s we e gene a ed, wi h a h eshold o
-log10(p- alue) > 4.5.
Geno yping
PACE p ime s
P ime iple s (Table1) o 3 selec ed SNPs in he newly
iden i ied QTL and o he SNP associa ed o paupe
we e designed by 3CR Bioscience (Essex, UK) and oli-
gonucleo ides we e acqui ed om IDT (Co al ille IA,
USA). Fo all ou loci, he p ime s we e designed based
on he sequence o he sense s and.
PACE assay
Fo he geno yping assays, we used he PACE mas e
mix (2x concen a ion, s anda d ROX le el) p o ided by
3CR Bioscience (Essex, UK). Each indi idual eac ion
Page 6 o 14Pa yi e al. BMC Plan Biology (2025) 25:945
consis ed o 5µl o he PACE mas e mix, 0.14µl o he
p ime pool and 5µl o ELGA wa e (VWS L d.; High
Wycombe, UK) o DNA empla e, isola ed om lea
samples ans e ed on o Qiagen FTA ca ds (QIAGEN
GmbH; Hilden, Ge many), p ocessed acco ding o he
manu ac u e `s ins uc ions. The cycling condi ions we e
se acco ding o he manu ac u e `s ins uc ions o he
analogously unc ioning Kompe i i e Allele Speci ic PCR
(KASP) assay (LGC G oup; Tedding on, UK) and an on
he CFX Opus 384 Real-Time PCR Sys em (Bio-Rad Lab-
o a o ies, Inc.; He cules CA, USA).
All 112 F2 indi iduals we e geno yped in 2023 using a
newly designed PACE ma ke agging he mos impo an
SNP associa ed o paupe [14], Lalb_Ch 18_12359687
and only e ospec i ely geno yped in 2024 upon he
a ailabili y o he PACE ma ke agging he Die a-asso-
cia ed swee ness. Samples om he F3 popula ion we e
geno yped wi h bo h ma ke s in 2024.
Da a analysis
All da a excep o he sequencing da a was analyzed and
isualized in R e sion 4.4.2 [27].
The pheno ypic da a (D agendo sco e) was ana-
lyzed in he F2 and F3 expe imen al popula ions
using Linea Mixed Models (LMM). The allelic com-
bina ions a loci Lalb_Ch 18_12359687 and Lalb_
Ch 05_6643621 we e included as ixed e ec , and
he ow he speci ic plan belonged o on he ield as
a andom e ec . The es ima ed ma ginal means om
he abo e-men ioned linea mixed model we e calcu-
la ed o es he s a is ical di e ence (p < 0.05) using
emmeans_1.10.7 [28] and we e used o assess he asso-
cia ion be ween he allelic s a es and he pheno ypic
alue in F2 and F3.
Simila ly, he o al alkaloid con en (sum o all QAs
p esen in a sample) and he a io be ween he concen-
a ions o 13α-hyd oxylupanine and lupanine pe swee
a ian we e analysed LMM using lme4_1.4–36 [29]. Fo
bo h esponse a iables we included he allelic combina-
ions as ixed, and he biological eplica e (i.e. se o sam-
ples belonging o allelic combina ions o he wo loci o
in e es ) as andom ac o . The es ima ed ma ginal means
om he abo e-men ioned linea mixed model we e cal-
cula ed o es he s a is ical di e ence (p < 0.05) using
emmeans_1.10.7 [28] and we e used o he ollowing
downs eam analysis.
Resul s
The pheno ypic seg ega ion in he F2 gene a ion con i ms
ano he de e minan o low QA con en
Wi h paupe known o seg ega e in he ‘Die a’ x ‘F ieda’
F2 popula ion by he geno yping a he causal mu a ion
SNP Lalb_Ch 18_12359687, he simples hypo hesis o
1: 3 (swee : bi e ) pheno ypic seg ega ion a io could be
ejec ed. Ins ead, he pheno ypic seg ega ion pa e n was
7: 9 (swee : bi e ), which coupled wi h obse a ion o he
geno ypic seg ega ion 1: 2: 1 a he paupe locus sugges s
he seg ega ion o a leas one addi ional low QA con e -
ing gene aside om paupe (Table2).
Assuming seg ega ion o wo independen ecessi e
genes, he pheno ypic sco es a e expec ed o seg ega e
in a a io o 7: 9 (swee : bi e ) in QA con en ( ollow-
ing Mendelian inhe i ance), which hypo hesis could no
be ejec ed (Table2), indica ing wo independen eces-
si e loci con e ing low QA con en wi h no seg ega ion
dis o ion in he F2 popula ion used o he BSA. The
assump ion o geno ypic seg ega ion o 1: 2: 1 could no
be ejec ed o ei he loci, u he suppo ing he inde-
pendence o hese de e minan s.
Iden i ica ion o a QTL associa ed o low QA con en
h ough BSA
A e he con i ma ion o he p esence o a de e mi-
nan independen o paupe o low QA con en in he
Table 1 PACE p ime s used in he s udy. Ta ge SNP locus on he ch omosome, he oligonucleo ide sequence o he wo o wa d
(F1 and F2) agging he wo possible alleles a he espec i e locus (unde lined), including he sequence o he X- and Y- Tails and he
e e se (R) p ime s o each iple , and mel ing empe a u es
Locus P ime Sequence (5’−3’) Tm (°C)
Lalb_Ch 05_6643621 F1 GAAGGTGACCAAGTTCATGCTAGACTATGTCATGTTTTTAAACATGTCAC64.2
F2 GAAGGTCGGAGTCAACGGATTGAGACTATGTCATGTTTTTAAACATGTCAA65.4
R GGCTTGAATAGGGTCCATGTGTTTAATAT 57.1
Lalb_Ch 05_6643864 F1 GAAGGTGACCAAGTTCATGCTAATCTATTATTGTCCTATTAAAACATTACGAG62.8
F2 GAAGGTCGGAGTCAACGGATTAATCTATTATTGTCCTATTAAAACATTACGAC63.3
RTATATTTTTTAGCTTAGGATGTTTTGGATT 52.3
Lalb_Ch 05_6645052 F1 GAAGGTGACCAAGTTCATGCTCGATAAGAGAGTGGGAGGACCAA67.7
F2 GAAGGTCGGAGTCAACGGATTGATAAGAGAGTGGGAGGACCAC67.8
R CGTGGCATGGACCCCAGTTATATAA 58.5
Lalb_Ch 18_12359687(paupe ) F1 GAAGGTGACCAAGTTCATGCTAAATGCTATCAGGATAGGGTCTATGT65.3
F2 GAAGGTCGGAGTCAACGGATTAAATGCTATCAGGATAGGGTCTATGA65.7
RCTTCCATTGTTCTTCCTCTATCTRCACTT 57.3
Page 7 o 14Pa yi e al. BMC Plan Biology (2025) 25:945
seg ega ing F2 popula ion, we iden i ied he genomic
egion esponsible o his uniden i ied de e minan o
low QA con en h ough BSA. F om he aw eads ang-
ing be ween 97 and 117 million pe bulk ( ead leng h o
150bp), 47–57 million emained a e quali y il e ing,
mapping and emo al o edundancies (Supplemen a y
Table S7).
The e we e 1.55million polymo phic SNPs (1.76mil-
lion be o e q40 il e ing) om he lupin b owse ha
we e polymo phic be ween he pa en s, epo ed by
bc ools, om which 514,982 emained a e il e ing ou
monomo phic, low quali y and low dep h a ian s. Fol-
lowing, 23,396 gene-ancho ed haplo ypes wi h an a e -
age o 21 SNPs and an a e age ead dep h o 302 pe
haplo ype we e cons uc ed. The selec i e sweeps analy-
sis esul ed in a clea peak on ch omosome 5 (Fig.2) ha
allows o di e en ia e he ‘swee ’ om he ‘bi e ’ bulk.
Candida e gene selec ion and ma ke de elopmen
Ten genes we e associa ed wi h haplo ypes showing sig-
ni ican (-log10(p- alue) > 4.5) di e ences in allele e-
quencies be ween swee and bi e bulks wi hin he en i e
QTL egion spanning om 5,799,140bp o 6,737,917bp
on ch omosome 5 (Supplemen a y Table S9). Al hough
none o hem could be di ec ly linked o he biosyn-
he ic pa hway, he gene Lalb_Ch 05g0222381 encoding a
“Pu a i e aminoacyl ans e ase” (a unc ion simila o he
gene esponsible o he paupe -con e ed low QA con-
en ) was chosen o molecula ma ke de elopmen . We
in es iga ed all indi idual SNPs wi hin as well as 1,000bp
up- and downs eam o his gene.
The e a e 54 polymo phic loci be ween ‘Die a’ and
‘F ieda’ in his egion. To de e mine whe he hese a e
also polymo phic wi hin o he accessions as well, we col-
lec ed he allelic s a es o bo h loci o all 39 sequenced
accessions om he whi e lupin genome b owse
(Supplemen a y Table S5). Ou o hese 54 SNPs, only
9 a e highly polymo phic wi hin he 39 sequenced
accessions [15]. Fu he mo e, only 3 loci a e consis-
en wi h he assumed alkaloid s a us o hese acces-
sions, Lalb_Ch 05_6643621, Lalb_Ch 05_6643864,
Lalb_Ch 05_6645052. These h ee we e es ed on he
alida ion panel o 34 accessions o u he con i m
hei associa ion o alkaloid s a us (Supplemen a y
Table S5), in compa ison o he ma ke - ai associa-
ion o he ma ke agging paupe . A he locus Lalb_
Ch 05_6643864, an old cul i a , Näh quell (desc ibed o
ca y he swee ness mu a ion nu icius) ca ied he allele
Table 2 Seg ega ion a ios and χ2 s a is ics in he pheno ype (D agendo sco e) and a he paupe , and Die a-associa ed locus o
he en i e F2 popula ion
Seg ega ing popula ion Expec ed a io Obse ed a io d χ2 p- alue
Pheno ypic seg ega ion
‘Die a’x
‘F ieda’
1:3 41:69 1 8.8364 0.00295 **
7:9 41:69 1 1.8753 0.1709
Geno ypic seg ega ion
‘Die a’
x
‘F ieda’
paupe seg ega ion 1:2:1 31:47:34 2 3.054 0.2172
Die a-associa ed seg ega ion 1:2:1 22:63:27 2 2.1964 0.3335
Fig. 2 Manha an plo o -log10(p- alue) (y-axis, h eshold se o -log10(p- alue) = 4.5), o di e ences in allele equencies o haplo ypes be ween wo
bulks o he F2 popula ion ep esen ing high and low QA le els. The x-axis indica es he posi ion o he haplo ypes in each o he 25 ch omosomes. A
zoomed-in plo is included o ch omosome 5 o de ail
Page 8 o 14Pa yi e al. BMC Plan Biology (2025) 25:945
associa ed o he Die a- ype swee ness, and he wild- ype
allele a he o he wo selec ed loci in he egion hus his
locus was no chosen o u he analysis. Simila ly, a he
locus Lalb_Ch 05_6645052, when compa ing wo di e -
en pheno ypically swee indi iduals o he cul i a ‘S a ’
(desc ibed o ca y he swee ness mu a ion exiguus) lead
o inconsis en esul s, hus we selec ed only he SNP
Lalb_Ch 05_6643621 o u he geno yping o he F3
seg ega ing popula ion. In e es ingly, F ench cul i a s
which belong o he panel o esequenced accessions,
namely ‘Magnus’, ‘O us’, ‘Luxe’ and ‘Clo is’ ca ied he
alleles associa ed o swee ness bo h o he paupe locus
and he newly desc ibed one on ch omosome 5.
Ma ke - ai associa ions e eal low QA s acked allele
ecombinan s
The ma ke - ai associa ion be ween he newly
desc ibed Lalb_Ch 05_6643621 SNP, he paupe causal
mu a ion SNP Lalb_Ch 18_12359687 and he pheno-
ypic sco e (D agendo sco e) was con i med o he
indi iduals in he F2 and F3 popula ions (Supplemen-
a y Figu e S1-S3), in compa ison o hei pa en al
lines. In F2, we we e able o s a is ically di e en ia e
be ween indi iduals ca ying he swee geno ype asso-
cia ed o paupe , and disc imina e o bi e indi iduals
based on he D agendo sco e (Supplemen a y Table
S2). Howe e , in F3, only paupe swee indi iduals,
and bi e indi iduals we e clea ly s a is ically g ouped
oge he based on he D agendo sco e only (Supple-
men a y Table S2, Supplemen a y Table S4). o ob ain
mo e nuanced obse a ions on he quan i a i e alue
o alkaloids, we pe o med GC/MS and GC/FID ana-
ly ical QA quan i ica ion in a subse o he F3 popula-
ion ep esen ing di e en combina ions o allelic s a es
a he paupe causal mu a ion SNP on ch omosome 18
and he locus Lalb_Ch 05_6643621 on ch omosome 5,
used o he ollowing downs eam analysis. This subse
included he wo pa en al lines and he mos impo an
allelic combina ions a bo h loci (Table3) and he co -
esponding es ima ed ma ginal means alues om he
pos -hoc analysis a e model i ing.
Compa ing he concen a ions o he o al alkaloid
con en (Fig.3), we obse ed a dec ease in he es i-
ma ed ma ginal means o he o al alkaloid con en
in all swee chemo ypes. The F3 plan s ca ying he
allele paupe as sole con ibu o o he low QA phe-
no ype (PAU) we e g ouped oge he wi h ‘F ieda’
(FRA), he pa en con e ing he paupe - ype swee -
ness (Fig.3). Simila ly, he F3 p ogeny plan s in which
he low QA pheno ype was con e ed by he pa -
en ‘Die a’ (DSW), we e g ouped oge he wi h he
pa en al geno ype (DIA), as well as he F3 indi iduals
he e ozygous o paupe and homozygous swee a
Lalb_Ch 05_6643621 (HET), hough wi h highe a e -
age QA con en compa ed o he plan s ha a e swee
because o paupe (PAU). The s acked allele ecombi-
nan s (STA) ha e a e y low alkaloid con en ( ange
aw da a 11–62.3 ppm), lowe compa ed o ei he
pa en o p ogeny, indica ing a clea addi i e e ec
o s acking wo low QA con e ing de e minan s.
S acked allele ecombinan s (STA) ep esen ed he
mos d as ic educ ion o 12% and 8% o he o al a e -
age alkaloid con en o he pa en al cul i a s, ‘Die a’
and ‘F ieda’, espec i ely. The o al es ima ed ma ginal
mean o he alkaloid con en was 22.8 ± 10.4 ppm d y
weigh in seeds o he s acked allele ecombinan s,
close o 10 imes lowe han he sugges ed h eshold
o human consump ion.
QA p o iles in he F3 popula ion highligh s he s acked
allele e ec
Based on he p esence o speci ic QAs in he di e en
allelic combina ions s udied, we obse ed dis inc QA
p o iles associa ed o di e en low-QA de e minan s and
hei combina ion. In he wild- ype, bi e pheno ype
(BIT) 14 QAs we e iden i ied (Fig.4).
Table 3 Subse o F3 indi iduals analyzed wi h GC/MS and GC/FID, hei allelic s a e, numbe o samples (n), es ima ed ma ginal
means o o al Quinolizidine alkaloid (QA) con en (in ppm) indica ing s a is ical g ouping, and s anda d e o (SE). BIT– bi e wild ype
p ogeny; FRA– ‘F ieda’ pa en , homozygous swee a paupe , wild ype a Lalb_Ch 05_6643621; DIA– ‘Die a’ pa en , wild ype a paupe ,
homozygous swee a Lalb_Ch 05_6643621; PAU– homozygous swee a paupe , wild ype a Lalb_Ch 05_6643621; DSW– wild ype a
paupe , homozygous swee a Lalb_Ch 05_6643621; HET– he e ozygous a paupe , homozygous swee a Lalb_Ch 05_6643621; STA–
homozygous swee a paupe , homozygous swee a Lalb_Ch 05_6643621
Code Allele nQA con en (ppm) SE (ppm)
Lalb_Ch 18_12359687 (paupe ) Lalb_Ch 05_6643621
BIT(bi e p ogeny) AA wild ype CC wild ype 4 9715.2 2284.8
DSW(Die a swee p ogeny) AA wild ype AA homozygous swee 10 296.4c15.7
HET TA he e ozygous AA homozygous swee 10 267.1c15.7
DIA(‘Die a’ pa en ) AA wild ype AA homozygous swee 10 257.4c15.7
PAU(paupe swee p ogeny) TT homozygous swee CC wild ype 7 171.7b18.5
FRA(‘F ieda’ pa en ) TT homozygous swee CC wild ype 10 176.2b15.7
STA(s acked allele ecombinan s) TT homozygous swee AA homozygous swee 23 22.8a10.4
Page 9 o 14Pa yi e al. BMC Plan Biology (2025) 25:945
Fig. 4 Gas ch oma og am o alkaloids om L. albus seeds. IS - in e nal s anda d (spa ein); 1 -ammodend ine; 2 - albine; 3 - e ahyd o hombi oline; 4 - an-
gus i oline; 5 - α-isolupanine; 6 - lupanine; 7 - N-me hylalbine; 8 - mul i lo ine; 9–17-oxolupanine; 10–13α-hyd oxylupanine; 11–13α-hyd oxymul i lo ine;
12–13α-angeloyloxylupanine; 13–13α- igloyloxylupanine; 14 − 13α-angeloyloxymul i lo ine
Fig. 3 Indi idual QA con en and composi ion, and o al QA con en (es ima ed ma ginal mean alue in ppm) o di e en ecombinan s in he F3
popula ion compa ed o he pa en al geno ypes. Values wi h he same le e a e no signi ican ly di e en a 0.05 con idence le el. BIT– bi e wild ype
p ogeny; FRA– ‘F ieda’ pa en , homozygous swee a paupe , wild ype a Lalb_Ch 05_6643621; DIA– ‘Die a’ pa en , wild ype a paupe , homozygous
swee a Lalb_Ch 05_6643621; PAU– homozygous swee a paupe , wild ype a Lalb_Ch 05_6643621; DSW– wild ype a paupe , homozygous swee a
Lalb_Ch 05_6643621; HET– he e ozygous a paupe , homozygous swee a Lalb_Ch 05_6643621; STA– homozygous swee a paupe , homozygous swee
a Lalb_Ch 05_6643621
