Gen om e d iver si ty

Scientists
Grant Positions
keilwagen, Jens (Saxony-anhalt, till 30.06.2012)
Seifert, Michael, dr. (Saxony-anhalt, till 30.04.2012)
Visiting Scientists/Scholars
Seifert, Michael, dr. (self-financed, 01.05.-31.12.2012)
Goals
Processing, visualization and interpretation of high-dimen-
sional biological data with modern machine learning methods.
Research Report
the “data inspection” (di) research group financed by the Min-
istry of culture of Saxony-anhalt, XP 3624hP/0606t analysed a 
broad spectrum of high-throughput biological data sets. our 
work began where very large and/or heterogeneous data can-
not be processed any more using standard statistical methods. 
the di group took a leading role in analyzing those sets with 
modern probabilistic modeling approaches in close coopera-
tion with biological groups inside and outside of the iPk. 
in addition to advanced data analyses for different iPk groups, 
the main focus of the di group was on the following areas:  
(1) computational genome comparison, (2) motif discovery, 
(3) reverse engineering, and (4) data mining in the cereals 
collection of the iPk Genebank.
(1) by combining the complementary strengths in epigenetics 
of dr. V. colot group (institut de biologie de l’ ecole normale 
Supérieure, France) with machine learning techniques of our 
group the aim was to identify newly inserted transposable 
elements (tes) . the huge data sets resulting from these ex-
periments was be analyzed with help of a new class of hidden 
Markov Models developed by M. Seifert (
http://www.jstacs.
de/index.php/PhhMM
 )  within the frame of a daad bilateral 
project.
(2) We developed the de-novo motif discovery tool dispom 
(http://www.jstacs.de/index.php/dispom) for finding differen-
tially abundant transcription factor binding sites that models 
existing positional preferences of binding sites and adjusts the 
length of the motif in the learning process. J. keilwagen in co-
operation with colleagues from Martin Luther university halle 
was able to show that the prediction performance of this tool 
is superior to existing tools for de-novo motif discovery. Finally, 
together with the group of i. Paponov from the albert Ludwigs 
university of Freiburg the tool was applied for discovering 
binding sites enriched in promoters of auxin-responsive genes. 
Research Group: Data Inspection 
(till 31 december 2012)
head:  dr. Swetlana Friedel

Abteilung Molekulare Genetik/
Department of Molecular Genetics
110
2013
a
ghaeepour
,  n., G. F
inak
, h. h
oos
, t.R. M
osmann
, R.R. b
rinkman
, R. 
G
ottardo
, R.h. S
cheuermann
, d. d
ougall
, a.h. k
hodabakhshi
, P. 
M
ah
, G. o
bermoser
, J. S
pidlen
, i. t
aylor
, S.a. W
uensch
, J. b
ram
-
son
, c. e
aves
, a.P. W
eng
, e.S. F
ortuno
  iii, k. h
o
, t.R. k
ollmann
, 
W. R
ogers
, S. 
de
 R
osa
, b. d
alai
, a. a
zad
, a. P
othen
, a. b
randes
, 
h. b
retschneider
, R. b
ruggner
, R. F
inck
, R. J
ia
, n. z
immerman
, M. 
L
inderman
, d. d
ill
, G. n
olan
, c. c
han
, F.e. k
hettabi
, k. o’n
eill
, M. 
c
hikina
, Y. G
e
, S. S
ealfon
, i. S
ugár
, a. G
upta
, P. S
hooshtari
, h. z
are
, 
P.L. 
de
 J
ager
, M. J
iang
, J. k
eilwagen
, J.M. M
aisog
, G. L
uta
, a.a. b
ar
-
bo
, P. M
ájek
, J. V
ilček
, t. M
anninen
, h. h
uttunen
, P. R
uusuvuori
, M. 
n
ykter
, G.J. M
c
L
achlan
, k. W
ang
, i. n
aim
, G. S
harma
, R. n
ikolic
, S. 
P
yne
, Y. Q
ian
, P. Q
iu
, J. Q
uinn
, a. R
oth
, P. M
eyer
, G. S
tolovitzky
, J. 
S
aez
-R
odriguez
, R. n
orel
, M. b
hattacharjee
, M. b
iehl
, P. b
ucher
, k. 
b
unte
, b. d
i
 c
amillo
, F. S
ambo
, t. S
anavia
, e. t
rifoglio
, G. t
offolo
, 
S.d. S
lavica
 d
imitrieva
, R. d
reos
, G. a
mbrosini
, J. G
rau
, i. G
rosse
, 
S. P
osch
, n. G
uex
, M. k
ursa
, W. R
udnicki
, b. L
iu
, M. M
aienschein
-
c
line
, P. S
chneider
, M. S
eifert
, M. S
trickert
 & J.M.G. V
ilar
: criti-
cal assessment of automated flow cytometry data analysis 
techniques. nat. Methods 10 (2013) 228-238.
& t. a
ltmann
: heterosis manifestation during early Arabidop-
sis seedling development is characterized by intermediate 
gene expression and enhanced metabolic activity in the 
hybrids. Plant J. 71 (2012) 669-683.
M
önke
, G., M. S
eifert
, J. k
eilwagen
, M. M
ohr
, i. G
rosse
, u. h
ähnel
, a. 
J
unker
, b. W
eisshaar
, u. c
onrad
, h. b
äumlein
 & L. a
ltschmied
: to-
wards  the identification and regulation of the Arabidopsis 
thaliana abi3-regulon. nucleic acids Res. 40 (2012) 8240-
8254.
S
eifert
, M., S. c
ortijo
, M. c
olome
-t
atche
, F. J
ohannes
, F. R
oudier
 & V. 
c
olot
: MediP-hMM: Genome-wide identification of distinct 
dna methylation states from high-density tiling arrays. bio-
informatics 28 (2012) 2930-2939.
S
eifert
, M., a. G
ohr
, M. S
trickert
 & i. G
rosse
: Parsimonious high-
er-order hidden Markov Models for improved array-cGh 
analysis with applications to Arabidopsis thaliana. PLoS 
comput. biol. 8 (2012) e1002286.
t
hiel
, J., d. R
iewe
, t. R
utten
, M. M
elzer
, S. F
riedel
, F. b
ollenbeck
, W. 
W
esch
 
ke
 & h. W
eber
: differentiation of endosperm transfer 
cells of barley – a comprehensive analysis at the micro-
scale. Plant J. 71 (2012) 639-655.
Fig. 33
a novel strategy for valorizing genetic diversity stored in genebanks. every year genebank curators select accessions for seed multiplication. these accessions are grown 
in field trials and their phenotypic observations are recorded during the growing period. annual ranking of these data is the first step in normalized rank product analysis 
which allows comparing accessions that have been cultivated in different years and under different conditions. this example visualizes a wheat sample cultivated four times 
since 1946. the histograms indicate the distributions for all plant samples cultivated in a given year. Missing histograms indicate missing phenotypic observations. based on 
normalized rank products, multi-trait optimization allows identifying plant samples with specific combinations of traits that can be utilized for targeted plant research and 
breeding. the black star and red point in the cube represent the best virtual and one real plant sample, respectively, that simultaneously have early flowering time, small plant 
height and high thousand grain weight. in the histograms of absolute trait values, the actual plant sample is indicated by red lines.

111
a
liyu
,  o.M., M. S
eifert
, J.M. c
orral
, J. F
uchs
 & t.F. S
harbel
: copy 
number variation in transcriptionally active regions of sex-
ual and apomictic Boechera demonstrates independently-
derived apomictic lineages. Plant cell 25 (2013) 3808-3823.
G
rau
, J., J. k
eilwagen
, a. G
ohr
, i.a. P
aponov
, S. P
osch
, M. S
eifert
, M. 
S
trickert
 & i. G
rosse
: diSPoM: a discriminative de-novo mo-
tif discovery tool based on the JStacS library. J. bioinform. 
comput. biol. 11 (2013) 1340006 (20 pages).
G
rau
, J., S. P
osch
, i. G
rosse
 & J. k
eilwagen
: a general approach for 
discriminative de novo motif discovery from high-through-
put data. nucleic acids Res. 41 (2013) e197.
G
ruber
, b.d., R.F.h. G
iehl
, S. F
riedel
 & n. 
von
 W
irén
: Plasticity of the 
arabidopsis root system under nutrient deficiencies. Plant 
Physiol. 163 (2013) 161-179.
L
ermontova
,  i., M. k
uhlmann
, S. F
riedel
, t. R
utten
, S. h
eckmann
, M. 
S
andmann
, d. d
emidov
, V. S
chubert
 & i. S
chubert
: Arabidopsis ki-
netochoRe nuLL
2
 is an upstream component for cenh3 
deposition at centromeres. Plant cell 25 (2013) 3389-3404.
M
ascher
, M., i. S
chubert
, u. S
cholz
 & S. F
riedel
: Patterns of nucleo-
tide asymmetries in plant and animal genomes. bioSystems 
111 (2013) 181-189.
W
eirauch
,  M.t., a. c
ote
, R. n
orel
, M. a
nnala
, Y. z
hao
, t.R. R
iley
, J. 
S
aez
-R
odriguez
, t. c
okelaer
, a. V
edenko
, S. t
alukder
, h.J. b
us
-
semaker
, M.d. Q
uaid
, M.L. b
ulyk
, G. S
tolovitzky
, t.R. h
ughes
, P. 
a
gius
, a. a
rvey
, P. b
ucher
, c.G. c
allan
 J
r
., c.W. c
hang
, c.-Y. c
hen
, 
Y.-S. c
hen
, Y.-W. c
hu
, J. G
rau
, i. G
rosse
, V. J
agannathan
, J. k
eilwa
-
gen
, S.M. k
iebasa
, J.b. k
inney
, h. k
lein
, M.b. k
ursa
, h. L
ähdesmäki
, 
k. L
aurila
, c. L
ei
, c. L
eslie
, c. L
inhart
, a. M
urugan
, a. M
yšičková
, 
W.S. n
oble
, M. n
ykter
, Y. o
renstein
, S. P
osch
, J. R
uan
, W.R. R
ud
-
nicki
, c.d. S
chmid
, R. S
hamir
, W.-k. S
ung
, M. V
ingron
 & z. z
hang
: 
evaluation of methods for modeling transcription factor se-
quence specificity. nat. biotechnol. 31 (2013) 126-134.
Books and Book Chapters
2012
k
ilian
, b., h. Ö
zkan
, S. S
haaf
, S. h
übner
, R.k. P
asam
, R. S
harma
, k. n
eu
-
mann
, W. W
eissgerber
, F.a. k
onovalov
, J. k
eilwagen
, S. F
riedel
, h. 
k
nüpffer
, M. 
von
 k
orff
, G. c
oupland
 & a. G
raner
: comparing 
genetic diversity within a crop and its wild progenitor: a 
case study for barley. in: M
axted
, n., M.e. d
ulloo
, b.V. F
ord
-
L
loyd
, L. F
rese
, J.M. i
riondo
 & M.a.a. P
inheiro
 
de
 c
arvalho
 (eds.): 
agrobiodiversity conservation: Securing the diversity of 
crop Wild Relatives and Landraces. cabi Publishing, Wal-
lingford (2012) 186-192.

Abteilung Molekulare Genetik/
Department of Molecular Genetics
112
fragments autocatalytically in a process called intein-mediated 
protein splicing. the transformed pro-vector constructs were 
shown to enable the production of male-sterile wheat lines by 
in planta assembly of the barnase protein fragments (M. Gils, 
k. kempe, M. Rubtsova). Male-sterility was found to be stable 
over several generations and under increased temperatures, 
without compromising female fertility. in order to achieve the 
positioning of the barnase gene fragments to allelic positions, 
the split gene system requires functional single-copy integra-
tions of the pro-locus. For this purpose, the expression of the 
split barnase fragments and the functionality of the intein sys-
tem were improved by modifying the basal pro-vectors (M. Gils, 
Fig. 34 b). a favourable effect resulted from the introduction 
of multiple GGGGS peptide linkers. in addition, the insertion of 
introns for intron-mediated enhancement (IME) of gene ex-
pression led to a significant improvement of the split-barnase 
system (k. kempe, M. Rubtsova, M. Gils). three different introns 
(from Petunia hybrida and Arabidopsis thaliana) were inserted 
into different positions of the c- and n-terminal coding regi-
ons and combined in a series of pro-vectors. as revealed from 
the transformation of 14 pro-vectors into approximately 3.000 
primary transformants, we were able to efficiently generate 
male-sterile wheat plants with a single-copy insertion of the 
pro-locus. Such lines are the basis material for further system 
development.
Collaborative activities
the intein technology used for the split barnase approach was 
also employed for the production of spider silk protein multi-
merisation in transgenic tobacco plants within the framework 
of collaboration with the research group Phytoantibodies (u. 
conrad).
a functional phic31-based recombination system was estab-
lished for barley in collaboration with the research group Plant 
Reproductive biology (J. kumlehn, e. kapusi).
analysis of the central metabolism of male-sterile transgenic 
plants and segregating non-transgenic plants revealed that the 
central metabolism of the transgenic plants was not altered 
in any tissue except the anthers (collaboration with the group 
heterosis, d. Riewe). thus, a proof of substantial equivalence 
could be delivered. Protocols for a more efficient production 
of doubled haploid wheat plants using a colchicine-free anther 
culture approach were developed in the frame of a collabora-
tive effort with Saaten-union biotec Gmbh, Gatersleben.
Research Group: Hybrid Wheat 
(till 31 december 2013)
head:  dr. Mario Gils
Scientists
Grant Positions
kempe, katja, dr. (bMbF)
Rubtsova, Myroslava, dr. (bMbF)
Goals
establishment of an efficient pollination control system for hy-
brid wheat breeding based on a transgenic “Split Gene System“.
Research Report
The “Split Gene System” for hybrid wheat production
Male sterility is constituted by a barnase gene (ribonuclease 
derived from Bacillus amyloliquefaciens) that is expressed in ta-
petum causing pollen ablation. in order to be able to reverse 
the male sterile phenotype in wheat after the hybrid cross, the 
barnase gene is split into separated complementary fragments 
and positioned at isoallelic positions (Fig. 34 a, p. 113). the po-
sitioning is implemented by transforming a precursor vector 
(“pro-vector”) harbouring the two tapetum-expressed barnase 
gene fragments and the subsequent site-specific deletion at 
the transferred dna during plant development. For this pur-
pose, the pro-vectors harbour recognition sites for a site-speci-
fic recombinase (phic31 integrase). the site-specific deletion 
reactions result in the production of two alternative derivative 
loci, with each producing inactive barnase precursor peptides. 
crossing of plants that carry the respective complementary loci 
with each other leads to progeny plants harbouring the two 
barnase gene fragments in allelic positions. these plants are 
male-sterile and can be used as the female crossing partners 
for the production of hybrid seed. the t1 hybrids are fertile as 
the barnase gene fragments segregate in the progeny. 
For the maintenance of the female crossing partner, the hetero-
zygous male sterile plant can be crossed to a homozygous line. 
the  barnase gene fragments conferring male sterility can be 
linked to an herbicide tolerance gene. thus, the heterozygous 
plants (females) can be selected by applying an herbicide. the 
system allows for mixed breeding of father and mother lines. 
in order to increase the stability of the “split barnase protein 
complex”, the barnase gene fragments were fused to intein 
sequences that, upon translation, covalently fuse the protein 

113
k
empe
,  k., M. R
ubtsova
, d. R
iewe
 & M. G
ils
: the production of  
male-sterile wheat plants through split barnase expression 
is promoted by the insertion of introns and flexible peptide 
linkers. transgenic Res. 22 (2013) 1089-1105.
R
ubtsova
,  M., h. G
nad
, M. M
elzer
, J. W
eyen
 & M. G
ils
: the auxins 
centrophenoxine and 2,4-d differ in their effects on non-
directly induced chromosome doubling in anther culture of 
wheat (T. aestivum L.). Plant biotechnol. Rep. 7 (2013) 247-
255.
Books and Book Chapters
2012
G
ils
,  M., M. R
ubtsova
 & k. k
empe
: Split-transgene expression in 
wheat. in: d
unwell
,  J.M. & a.c. W
etten
 (eds.): transgenic 
plants: Methods and Protocols (Methods Mol. biol. 847). 
humana Press, c/o Springer Science+business Media, LLc, 
new York (2012) 123-135.
Publications
Peer Reviewed Papers
2012
k
apusi
, e., k. k
empe
, M. R
ubtsova
, J. k
umlehn
 & M. G
ils
: phic31 inte-
grase-mediated site-specific recombination in barley. PLoS 
one 7 (2012) e45353.
2013
G
ils
,  M., k. k
empe
, a. b
oudichevskaia
, R. J
erchel
, d. P
escianschi
, R. 
S
chmidt
, M. k
irchhoff
 & R. S
chachschneider
: Quantitative as-
sessment of wheat pollen shed by digital image analysis 
of trapped airborne pollen grains. adv. crop Sci. techn. 2 
(2013) 119.
h
auptmann
, V., n. W
eichert
, M. M
enzel
, d. k
noch
, n. P
aege
, J. S
cheller
, 
u. S
pohn
, u. c
onrad
 & M. G
ils
: native-sized spider silk prote-
ins synthesized in planta via intein-based multimerization. 
transgenic Res. 22 (2013) 369-377.
Fig. 34
a) the split-gene approach for hybrid seed production. t, tapetum-specific promoter. b) design of transformation vectors that are used for the production of male-sterile 
wheat plants. c) Pca analysis comparing the central metabolism of leaf and anther tissue isolated from male-sterile and segregating non-transgenic plants (M. Gils, k. kempe, 
d. Riewe).

Abteilung Molekulare Genetik/
Department of Molecular Genetics
114
Other Papers
2013
G
ils
,  M., M. R
ubtsova
 & k. k
empe
: come together: intein-media-
ted protein ligation in transgenic wheat. Proc. “Genetic 
transformation technologies (plants and animals) – Plant 
Genetics and breeding technologies – Plant diseases and 
Resistance Mechanisms” Vienna. Medimond international 
Proceedings (2013) 27-32.
Patents
2012
g
ils
, M., u. c
onrad
 & t
ran
 t
rong
 h
oang
: Methods of producing 
and purifying polymeric proteins in transgenic plants. eP 
2518081, Veröffentlichung: 31.10.2012, iPk-nr. 2011/07.

115
in the frame of the German Plant Phenotyping network (dPPn), 
the development of an approach to detect new organ-specific 
traits has been started. based on a corner detection algorithm, 
our system enables us to detect leaf-tips of maize plants  
(J.-M. Pape; Fig. 35 a and b, p. 116). now we are able to extract 
multiple geometric and color-related features. additionally, 
we im proved our segmentation approach to classify plant 
images based on predefined color classes, using the k-means 
algorithm. in preparation of the dPPn root project, a combined 
analysis of root and shoot data was performed. Statistical 
approaches such as canonical correlation and Pca (principal 
component analysis) were used (d. chen).
Within the european Plant Phenotyping network (ePPn) and 
in collaboration with the partners from the heterosis group, 
three test-experiments using artificial plants were performed. 
the common results were discussed within the consortium. to 
exchange datasets in a common format, the iSa-tab framework 
for metadata tracking was adopted. in the future, our iaP soft-
ware will incorporate this approved standard (d. chen). 
For the oPtiMaL project, several new phenotypic traits of  
maize plants can be detected. For example, a leaf-curling 
analysis based on the fast Fourier transform (FFt) has been 
developed. this new approach allowed us to characterize 
the frequency and amplitude of the leaf curling. a new post-
processing pipeline is used to interpret phenotypic datasets 
(d. chen). the pipeline includes several well-established 
statistical approaches, such as Pca and anoVa (analysis of 
variance). Working together with the data inspection group and 
the Genome diversity group (dept. Genebank), an approach 
to model plant growth was implemented (c. klukas, d. chen;  
Fig. 35 c - e). the post-processing pipeline not only works on 
high-throughput phenotyping datasets, but also works on field 
data. based on the pipeline, manually measured phenotypic 
data from field and greenhouse were analysed in collaboration 
with the Plant architecture group (dept. Genebank). 
in two collaboration projects with Prof. chen from zhejiang 
university, hangzhou, china (supported by the Federal office 
for agriculture and Food and the Robert bosch Stiftung), new 
extensions for our analysis framework iaP were implemented, 
to more flexibly support data transport and experiment stor-
age (c. klukas, d. chen). it is now possible to store and jointly 
access numeric data such as metabolite-, protein- or gene-ex-
pression-datasets. in the same way it became possible to ac-
cess binary data such as images from high-throughput experi-
ments in user-defined storage locations. Supported data trans-