Proceedings of the Workshop on Annotating and Reasoning about Time and Events, pages 46–53,
Sydney, July 2006. c©2006 Association for Computational Linguistics
Marking Time in Developmental Biology: Annotating Developmental
Events and their Links with Molecular Events
Gail Sinclair
School of Informatics
University of Edinburgh
Edinburgh EH8 9LW
c.g.sinclair@ed.ac.uk
Bonnie Webber
School of Informatics
University of Edinburgh
Edinburgh EH8 9LW
bonnie@inf.ed.ac.uk
Duncan Davidson
MRC Human Genetics Unit
Western General Hospital
Edinburgh EH4 2XU
Duncan.Davidson@hgu.mrc.ac.uk
Abstract
Current research in developmental biology
aims to link developmental genetic path-
ways with the processes going on at cel-
lular and tissue level. Normal processes
will only take place under speci c sequen-
tial conditions at the level of the pathways.
Disrupting or altering pathways may mean
disrupted or altered development.
This paper is part of a larger work explor-
ing methods of detecting and extracting in-
formation on developmental events from
free text and on their relations in space and
time.
1 Introduction
Most relation extraction work to date on biomedi-
cal articles has focused on genetic and protein in-
teractions, e.g. the extraction of the fact that ex-
pression of Gene A has an effect on the expression
of Gene B. However where genetic interactions are
tissue- or stage-speci c, the conditions that govern
the types of interactions often depend on where in
the body the interaction is happening (space) and
at what stage of life/development (time).
For genetic pathways involved in development,
it is critical to link what is happening at the molec-
ular level to changes in the developing tissues,
usually described in terms of processes such as
tubulogenesis and epithelialization (both involved
in the development of the kidney) and where they
are happening.
The processes themselves are usually linked
to stages rather than precise time points and
spans like  6.15pm EST ,  March 3 ,  last year .
Within the developmental mouse community,
there are at least two different ways of specifying
the developmental stage of an embryo - Theiler
stages (TS), and days post coitum/embryonic
day (d.p.c./E). However, these cannot be simply
mapped to one another as can days, weeks and
years. Embryonic days are real time stages in-
dependent of the state of the embryo and dated
from an assumption about when (approximately)
the relevant coitus must have taken place, while
Theiler stages are relative stages dependent on the
processes an embryo is undergoing.
Developmental stages can be also be referred to
implicitly, by the state of the embryo or the pro-
cesses currently taking place within it. This is be-
cause during development, tissues form, change,
merge or even disappear. So if the embryo is un-
dergoing tubulogenesis, one can assume that its
developmental stage is (loosely) somewhere be-
tween TS20 and birth. If the text refers to induced
mesenchyme during a description of tubulogene-
sis, one can assume that this change in the mes-
enchyme is the (normal) consequence of the Wolf-
fian duct invading the metanephric mesenchyme.
The invasion is known to occur around 10.5 d.p.c
so the induced mesenchyme must come into exis-
tence soon after this time.
Temporal links between developmental events
may be indicated explicitly (e.g. first, a tubule de-
velops into a comma-shaped body, which then de-
velops into an S-shaped body), but they are more
likely to be indicated implicitly by their order-
ing in the text and by associative (or  bridging )
anaphora where the anaphor refers to the result of
a previously mentioned process, e.g. the induction
of metanephric mesenchyme as one event, and a
subsequent mention of induced mesenchyme (an
 associative or  bridging reference) within an-
other event, suggesting the former event occurred
before the latter.
46
Figure 1: Partial genetic pathway for early kid-
ney morphogenesis. The arrows show directed in-
teractions between genes that are required for the
speci ed processes. E.g Pax2 interacts with (acti-
vates) Six2 which, together with Sall1 and Wt1,
is required for differentiation of the mesenchy-
mal cells in the metanephric mesenchyme. Image
taken from (Ribes et al., 2003).
This work on linking molecular and develop-
mental events mentioned in text on development
is also meant to deal with the problem that no
one article ever fully describes a topic. The par-
tial genetic interaction network in Figure 1 has
been built from several different studies and not
determined from just a single experiment. So
not only does the information within one article
need to be mined for useful information - the in-
formation across articles needs to be associated
with each other with respect to temporal, spatial
and experimental grounding. Eccles et al. (2002)
states that Pax2 is required for differentiation of
mesenchymal cells during kidney morphogenesis,
while Sajithlal et al. (2005) states that Eya1 is re-
quired. However these two results by themselves
do not help us determine whether these require-
ments are independent of one another or whether
they are required at different stages or in different
parts of metanephric mesenchyme or whether the
two genes interact. The conditions involved in the
experiments, most importantly the temporal con-
ditions, can help to link the two events.
This work aims to develop methods for extract-
ing information from text that will ground ge-
netic pathways (molecular events) with regard to
tissue location, developmental process and stage
of embryonic development - that is, their spatio-
temporal context. The task at hand is to recognise
how biologists write about developmental events
and then adapt existing or formulate new natu-
ral language processing techniques to extract these
events and temporally relate them to each other.
The resultant information can then be used both
for database curation purposes and for visualisa-
tion, i.e. to enrich pathway diagrams such as Fig-
ure 1, with information such as when and where
the interactions take place, what type of inter-
actions are involved (physical, activation, inhibi-
tion), the origin of this information and other as-
sociated information.
2 Notions of Time
As previously mentioned, there are different ways
of calibrating for developmental stages, and they
cannot simply be mapped to one another. The two
most common stage notations for mouse develop-
ment are Theiler stages, TS, and Embryonic days,
E (equivalent to days post coitum, d.p.c.). The lat-
ter are self explanatory in that they denote the 24
hour day and can be considered real-time staging.
47
The convention was originally that E11 would rep-
resent the 24 hour period of the 11th day. It is,
however, now common to  nd E11.5 representing
the same time period, but this is merely a change
in convention due to standard practices of experi-
mentation.
A Theiler stage on the other hand represents
a non- xed relative time period de ned by the
progress of development rather than directly in
terms of the passage of time. Theiler Stages
(Theiler, 1989) divide mouse development into 26
prenatal and 2 postnatal stages. In general, Theiler
used external features that can be directly assessed
by visual inspection of the live embryo as devel-
opmental landmarks to de ne stages. The Edin-
burgh Mouse Atlas Project (EMAP)1 uses Theiler
stages to organise anatomical terms in their Mouse
Atlas Nomenclature (MAN). EMAP gives a brief
description of each Theiler stage with TS25 as an
example as follows:
Skin wrinkled
The skin has thickened and formed wrinkles and
the subcutaneous veins are less visible. The  n-
gers and toes have become parallel and the um-
bilical hernia has disappeared. The eyelids have
fused. Whiskers are just visible.
Absent: ear extending over auditory meatus, long
whiskers.
An embryo is in TS25 at approximately 17 d.p.c.
As can be seen in Figure 2, an embryo at E11
could be considered in Theiler stage 17, 18 or 19,
i.e. Theiler stages can overlap one another with
respect to Embryonic day. Indeed, here, TS17 can
fully encompass TS18 in the dpc timeline.
The development of internal structures is ap-
proximately correlated with external develop-
ments, so except for  ne temporal differences,
the Theiler stages can be assumed to apply to the
whole embryo. Theiler stages provide only gross
temporal resolution of developmental events, and
the development of internal structures often take
place within the boundaries of one of these stages
or overlapping stage boundaries. Thus, internal
developmental processes can also have their own
 ner relative timeline or staging.
There is no ontology or reference book that
comprehensively speci es this  ner staging and
the knowledge of the biologist as the reader of ar-
1Edinburgh Mouse Atlas Project -
http://genex.hgu.mrc.ac.uk/
Figure 2: Graphic of kidney morphogenesis an-
notated with the two standard staging notations
for mouse development. At E10.5 the Wolf an
duct invades the metanephric mesenchyme form-
ing the ureteric bud around E11. The bud then
branches around E11.5 and continues to do so un-
til birth, forming the ultimate functional units of
the kidney - the nephrons. TS = Theiler Stage, E
= Embryonic day/dpc. This image is adapted from
http://www.sciencemuseum.org.uk/
ticles is relied upon. This work will contribute to
making this deeper staging criteria explicit.
3 Annotation
3.1 Event Classification
As a  rst step, a Gold Standard corpus of 988 sen-
tences was developed with each sentence being
classi ed as containing the description of a devel-
opmental and/or molecular event or not. 385 sen-
tences were classi ed as positive, with 603 neg-
ative. Named entities within all these sentences
were also annotated. Among these element types
were stage, process and tissue. A Naive Bayes
automatic classi er for sentence classi cation was
developed using this Gold Standard resulting in
a balanced F-score of 72.3% for event classi ca-
tion. (A manual rule-based approach resulted in
an F-score of 86.6%, but this has yet to be fully
investigated for automation. Guessing positive for
all sentences would give a balanced F-score of
58.4%)
3.2 Event Specifications
Two event types are of interest in this work -
molecular and tissue events. The former involve
the action (and possible effect) of molecules dur-
48
ing development and the latter involves the devel-
opment of the tissues themselves. A description of
an event can be expected to contain the following
elements:
a0 molecular or tissue event type (e.g. expres-
sion, inhibition)
a0 stage or temporal expression (e.g. after X,
subsequent to X, E11)
a0 at least one of
– molecule name, anatomical term, bio-
logical process term
The informational elements included within an
event description can then be used to relate events
to each other. Speci cally, processes involve
known tissues and are known to happen during
certain stages, just as the relative order of pro-
cesses, tissue formations and stages are known.
While an initial speci cation of an event may
be associated with a single sentence, clause or
phrase, not all the elements of relevance to this
work may be speci ed there. In particular, an in-
formational element of the event may be explic-
itly and fully stated in this initial event speci ca-
tion, or it may be underspeci ed or it may be miss-
ing. For those that are underspeci ed or missing,
background knowledge about other elements and
events may need to be taken into consideration in
order for them to be fully resolved (see Section
4.2).
The following is a straightforward example
where the given sentence speci es all the main el-
ements required for a molecular event.
1. At E11, the integrin α8 subunit was expressed
throughout the mesenchyme of the nephro-
genic cord.
a0 Molecular Event : expression
a0 molecule name: integrin α8
a0 anatomical term: mesenchyme of the
nephrogenic cord
a0 stage: E11
Example 2 shows that a single sentence may
specify more than one event.
2. Prior to formation of the ureteric bud,
no α8 expression was evident within the
mesenchyme that separates the urogenital
ridge from the metanephric mesenchyme and
within the metanephric mesenchyme itself.
a0 EVENT-0
– Tissue Event : formation of anatom-
ical term
– anatomical term: ureteric bud
– stage/temporal expression = missing
a0 EVENT-1
– Molecular Event: absence of expres-
sion
– molecule name: α8
– anatomical term: mesenchyme that
separates the urogenital ridge from
the metanephric mesenchyme
– temporal expression:Prior to
EVENT-0
a0 EVENT-2
– Molecular Event: absence of expres-
sion
– molecule name: α8
– anatomical term: metanephric mes-
enchyme
– temporal expression: Prior to
EVENT-0
EVENT-0 is not the focus of this sentence, but
rather a reference event. Its attributes need to be
recorded so that the stage of the other events can
be determined.
TimeML (Pustejovsky et al., 2004) is a spec-
i cation language designed for the annotation
of temporal and event information. Although
TimeML is not currently being used as a method
of representation for this work, Example 1 above
could be represented as follows:
a1 SIGNAL sid=“s1” type=“temporal”
a2
At
a1 /SIGNAL
a2
a1 TIMEX tid=“t1” type=“STAGE”
value=“E11” a2
E11
a1 /TIMEX
a2
the integrin α8
a1 EVENT eid=“e1” class=“molecular”
a2
was expressed
a1 /EVENT
a2
throughout the mesenchyme of the
a1 SIGNAL sid=“s2” type=“tissue”
a2
nephrogenic cord
a1 /SIGNAL
a2
nephrogenic cord can be considered a signal of
type  tissue as it does not exist throughout the
49
whole of development and so can indicate or rule
out time periods for this event description.
3.3 Event Time-Stamping
The relative timing of any biological processes
mentioned in the event descriptions  rst needs to
be determined before we can work out when the
actual events described are taking place.
Schilder and Habel (2001) looked beyond the
core temporal expressions and into prepositional
phrases that contained temporal relations, i.e. be-
fore, during, etc and introduced the notion of noun
phrases as event-denoting expressions. An event
that is described as occurring  after the election 
does not have an explicit time-stamp attached to it,
but the knowledge about the timing of the election
mentioned gives the reader a notion of when in ab-
solute time the event occurred. This is similar to
Example 2 above where Event-0 is the reference
event, thus biological processes can be considered
event-denoting expressions.
While Schilder and Habel rely on prepositional
phrases to designate their event-denoting noun
phrases, for this work propositional phrases are
not necessarily required. The mention of a noun
phrase by itelf may be enough. In developmen-
tal biology, tissues may only be extant for a lim-
ited period before they form into some other tissue
and these can also be used as event-denoting ex-
pressions - for example, comma-shaped bodies are
structures within the developing kidney that are
only in existence for a relatively short time period -
before the existence of the S-shaped bodies and af-
ter epithelialization. Therefore the mention of tis-
sues as well as processes can help to pinpoint the
timing of the event being described. While they
may not ultimately bring us to the exact stage the
event is occurring in, it can at least rule out some
spans of time. We discuss this further in Section
4.2.
In order for events to be linked to one another,
it is necessary to uniquely index each event and its
elements. Mapping across indices will be utilised
so that known relationships between elements can
be represented. For example, E10 comes before
E12, tubulogenesis occurs during kidney morpho-
genesis, and the proximal tubule is part of the
nephron.
Of the elements types listed in Section 3.2, only
the molecule element cannot be used to resolve
developmental stage while tissue, process, stage
and, of course, temporal expression can. Other el-
ements are also of interest to the biologist and inte-
gral to development and molecular function, how-
ever they are not of use in the grounding of events
in time.
4 Initial Investigations
This section demonstrates that one must look be-
yond the sentence in order to resolve the temporal
aspects of events.
4.1 Evidence for Developmental Stage
Evidence suf cient to resolve developmental stage
can come from many places. 314 positive sen-
tences from the Gold Standard corpus and their
context were examined, and the evidence required
to resolve developmental stage for each of the
events mentioned there was determined as shown
in Table 1.
As can be seen from the table, only 48 out of
the 314 event sentences (i.e. 15%) have the de-
velopmental stage in which the event is occurring
explicitly stated in the given sentence, (e.g. Ex-
ample 1 in Section 3). So other means need to be
explored in order to ground events with respect to
developmental stage. An event sentence may be a
continuation of a topic, and so the speci c devel-
opmental stage involved may well be stated in the
immediately surrounding or related text.
Information in the immediately surrounding
text (rows labelled Following Sentence, Previous
Sentence and Current Paragraph) resolves the de-
velopmental stage of the event in 64 cases (i.e.
21%). This most commonly occurs by looking for
the immediately previously mentioned stage, and
in one case the next encountered stage.
Event sentences also often refer to  gures, and
so the stage being described in the caption (i.e.
legend) of the referenced  gure will often be the
same as the one relevant to the sentence. (This was
true of all sentences looked at that referenced a  g-
ure.) Figures, however, are generally only found in
the Results sections and so this type of evidence is
not often going to be of use for sentences found in
other sections of an article.
Similarly, events can be described within the
 gure legends themselves. The concise and simple
way in which legends are generally written mean
that the explicit stage is commonly referred to, and
so stage can be resolved using this referenced in-
formation(43 out of 47 cases, i.e. 91%).
50
Source of Evidence Abstract Introduction Results Discussion Methods Totals
Time Irrelevant 7 12 22 23 1 65
Prior Knowledge 17 33 31 45 0 126
Following Sentence 0 0 1 0 0 1
Previous Sentence 0 0 7 0 0 7
Current Paragraph 0 0 18 1 0 19
Reference to Figure 0 0 38 0 0 38
Within Fig Legend 0 0 43 0 0 43
(time not resolved) 0 3 1 0 0 4
Explicitly Stated 0 1 41 5 1 48
(not relevant) 0 0 1 0 0 1
Totals 24 49 165 74 2 314
Table 1: Location and type of evidence suf cient to resolve developmental stage in sentences. Time
Irrelevant indicates that the event being described is not time critical, i.e. event is a constant over devel-
opmental timeline, or end result. Prior knowledge means temporal information other than that found in
the current paragraph but associated with current event such as tissue and process is required for temporal
resolution. This may be found in the current article or from previously curated information (assuming
accurate terminology mapping.) Text from outside the current paragraph cannot be relied upon to be rel-
evant to the current sentence without additional information. time not resolved means the stage could not
be pinpointed using the  gure legend. not relevant indicates that although an explicit stage was referred
to within the sentence, this was not relevant to the event being described, e.g. event and stage in different
clauses of the sentence.
Table 2 shows a similar table to Table 1, but
deals only with those sentences found within  g-
ure legends. It shows where within the  gure leg-
end the required evidence for developmental stage
can be found. As can be seen, in 80% of these
cases the relevant developmental stage can be as-
certained directly from the legend. It should be
noted that  gure legends in biological articles tend
to be much lengthier than those from NLP articles.
In 21% of the event sentences, a speci c devel-
opmental stage is not relevant to the fact being de-
scribed ( rst row of Table 1), e.g. the kidneys of
the double mutants were located more caudal and
medial than normal. This sentence is describing
an end result, i.e. an affected or normal kidney
at birth (although this could, of course, be con-
sidered a developmental stage.) Alternatively, the
time-irrelevant event being described could be a
non-event, e.g. the fact that a gene is never ex-
pressed in a particular tissue. Similarly, this could
be considered as the developmental stage range
from conception to birth.
The signi cantly small proportion of event sen-
tences located in Abstracts (24 of 314 total event
sentences, less than 8%) demonstrates the need to
use full text. Even where an event is described
within an Abstract, it is rarely accompanied by
associated processes or tissues speci c enough to
suggest the stage of development never mind an
explicit timestamp, as it is, by necessity, only gen-
erally describing the whole article. The majority
of BioNLP work is being done with the use of Ab-
stracts only. This is because of their relative ease
of access compared with full text, but methods de-
veloped using Abstracts only will not necessarily
be as effective when applied to full text.
As can be seen, the majority of temporally-
underspeci ed event sentences are situated in the
Results section of the articles. Indeed, this is
the section where most event sentences are to be
found. This work is initially focussing on event
descriptions found in Results sections of articles as
these will focus on the work done by the authors
and their  ndings and will not generally include
modality in the event descriptions as Introduction
and Discussion sections might. As shown above,
the Methods section rarely contains event descrip-
tions and when they do they are usually about what
the experiment aims to show and so this should be
repeated in the Results section.
4.2 Prior Knowledge
As mentioned earlier, if none of the above sources
reveal the relevant stage of an event, then other el-
ements within the sentence, such as tissue or pro-
cess, need to be looked at so that prior knowledge
51
Source of Evidence Figure Legends
Time Irrelevant 4
Prior Knowledge 4
Following Sentence 0
Previous Sentence 14
Current Paragraph 13
Explicitly Stated 11
Total 47
Table 2: Location and type of evidence suf cient to resolve developmental stage in sentences within
 gure legends. Rows as in Table 1, with Current Paragraph being equal to the whole of the legend.
about those elements can be exploited for devel-
opmental stage to be resolved. For example, given
the sentence
Prior to formation of the ureteric bud,
no α8 expression was evident within the
mesenchyme that separates the urogen-
ital ridge from the metanephric mes-
enchyme and within the metanephric
mesenchyme itself.
the developmental stage can be resolved if we
know when the ureteric bud forms (TS17/E10.5).
It could also be the case that the other tissues
or processes mentioned have a speci c lifetime
within development and these could help to fur-
ther pinpoint the timeline involved for the lack of
α8 expression. For example,
Pax2 was initiating in the metanephric
mesenchyme undergoing induction.
It is not so straightforward to assign a stage here,
since the mesenchyme is constantly being induced
from E11 (TS18) until birth (TS26), but we have
at least discounted E1-E10 (TS1-TS17) as relevant
stages.
Resources such as the Mouse Atlas Nomencla-
ture (MAN) (Ringwald et al., 1994) will provide
the initial prior knowledge in order to resolve de-
velopmental stage of events. This describes the
different stages of development and the tissues in
evidence at each stage, giving what is known as the
abstract mouse. From this abstract mouse, we can
ascertain the normal stage ranges where tissues ex-
ist and use this knowledge for temporal resolution,
taking care not to assume that tissues do not neces-
sarily exist within the same stage range in mutant
mice than in wild-type. The prior knowledge data-
bank can be recursively added to with facts from
events already extracted from papers for use in fur-
ther event extraction and their anchoring in time.
5 Future Work
5.1 Term Normalisation
There is no point extracting events descriptions if
we cannot relate the events and their elements to
each other. The event-denoting expressions iden-
ti ed need to be normalised so that it can be recog-
nised when two terms are referring to the same el-
ement.
Inconsistent terminology in the biomedical  eld
is a known problem (Sinclair et al., 2002). One
gene can have several names (synonymy) just as
the same name can be used for more than one gene
(homonymy). Very often the synonyms bear no re-
lation to one another since they were perhaps con-
currently discovered in different laboratories and
named. For example, the gene insomnia can also
be known as cheap date, since experiments found
that organisms without this gene have a tendency
to fall asleep and are particularly susceptible to
alcohol. The same anatomical part can also be
referred to by different terms, e.g. the Wolffian
duct is also known as the nephric duct, and the
metanephros is another name for the kidney. There
is also a lineage issue, where a tissue with one
name (or perhaps more) develops into something
with another name (e.g. the intermediate meso-
derm gives rise to both the Wolffian duct and the
metanephric mesenchyme which in turn both de-
velop into the metanephros. The MAN includes
this type of information.
Term normalisation is particularly important for
the process and tissue elements. If these terms
are not normalised, temporal knowledge about the
terms may not be exploited and it may not be de-
termined that events involving them are linked.
52
5.2 Event Elements
If the elements required to fully describe an event
are explicitly stated within a simple sentence, then
temporal grounding will be straightforward. How-
ever, this is unlikely to often be the case. More
complex sentences will dictate the need for de-
pendency relations to be determined so that each
event’s elements can be identi ed. Methods for
dealing with missing or underspeci ed elements
that are not resolved within the event description
itself will be investigated.
A naive approach will  rst be investigated to
 ll these gaps:  nd the closest appropriate ele-
ment in the previous context (varying the size of
the window for how far back to look, such as cur-
rent paragraph or last 3 sentences). An error anal-
ysis on this simple method will help to guide the
amount of further work necessary to achieve equal
success across all elements. For those elements
that this method is ineffective, other methods will
be developed incorporating features such as sen-
sitivity to syntax, event type and location within
article. Similarly, it will be established whether
different techniques are required for missing infor-
mation than for underspeci ed information. They
will  rst be treated in the same manner with anal-
ysis determining whether they should be treated
differently.
6 Conclusion
This ongoing work has shown the importance of
relative time lines in order to link events to one
another. The identi cation of event elements and
their normalisation will then form a basis for rea-
soning over these elements with regards to  rst
time-stamping of events and then temporally relat-
ing the events. The aim of many BioNLP studies
is ultimately to reason over extracted events and,
as such, the relative timing of these events is cru-
cial. For example, if we know
1. tissue X is transformed into tissue Y at stage S
and
2. molecule M is expressed in X at stage S-1,
then it can be reasoned that event 2 has an im-
pact on event 1. This reasoning can be made more
successful if we know as much about the events
as possible, not just that tissue Y is formed and
molecule M is expressed.
It has also been demonstrated that we not only
need to look beyond the sentence level for tempo-
ral resolution but also beyond the article in order to
replicate the reader’s assumed level of background
knowledge.

References
J. F. Allen, Towards a general theory of action and
time, Arti cial Intelligence, vol 23, pp 123-154,
1984.
M. R. Eccles, S. He, M. Legge, R. Kumar, J. Fox, C.
Zhou, M. French and R. W. Tsai, .PAX genes in
development and disease: the role of PAX2 in uro-
genital tract development. Int J Dev Biol, vol 46, no
4, pp 535-44, 2002.
J. Pustejovsky, I. Mani, L. Belanger, B. Bogurev, B.
Knippen, J. Littman, A. Rumshisky, A. See, S.
Symonen, J. Van Guilder, L. Van Guilder and M.
Verhagen, The Speci cation Langeuage TimeML.
in The Language of Time: A Reader, Oxford Uni-
versity Press, 2004.
D. Ribes, E. Fischer, A. Calmont and J. Rossert, Tran-
scriptional Control of Epithelial Differentiation dur-
ing Kidney Development. J Am Soc Nephrol, vol
14, pp S9-S15, 2003.
M. Ringwald, R. A. Baldock, J. Bard, M. H. Kaufman,
J. T. Eppig, J. E. Richardson, J. H. Nadeau and D.
Davidson, A database for mouse development. Sci-
ence, vol 265, pp 2033-2034, 1994.
G. Sajithlal, D. Zou, D. Silvius and P. X. Xu, Eya 1 acts
as a critical regulator for specifying the metanephric
mesenchyme. Dev Biol, vol 284, no 2, pp 323-36,
2005.
F. Schilder and C. Habel, From Temporal Expres-
sions to Temporal Information: Semantic Tagging
of News Message., in Proceedings of the ACL 2001
Workshop on Temporal and Spatial Information Pro-
cessing, Toulouse, France, pp 88-95.
G. Sinclair, B. Webber and D. Davidson, Enhanced
Natural Language Access to Anatomically Indexed
Data. in Proceedings of the ACL 2002 Workshop
on Natural Language Processing in the Biomedical
Domain, Philadelphia, pp 45-52.
K. Theiler, The House Mouse. Atlas of Embryonic De-
velopment. Springer Verlag New York, 1989.
