Hedging, Fact from claim, Science as mythology

Hedging erosion as a
mechanism for fact creation:
a small discourse analysis of
the Bourne corpus

Anita de Waard
Elsevier Labs/U Utrecht/UToronto

Single paper: Implication is hedged


(1) Both seminomas (2) b. the detection of
and the EC component miR-371-3 merely
of nonseminomas share reflects its expression
features with ES cells. pattern in ES cells,



(2) a. To exclude that



(2) a. To exclude that

(2) c. we tested by RPA (3) a. In many of the miR-371-3
miR-302a-d, another ES expressing seminomas and
cells-specific miRNA cluster nonseminomas, miR-302a-d was
(Suh et al, 2004). undetectable (Figs S7 and S8),



(2) a. To exclude that (3) b. suggesting



(1) Both seminomas (2) b. the detection of (3) c. miR-371-3
and the EC component miR-371-3 merely expression is a
of nonseminomas share reflects its expression selective event during
features with ES cells. pattern in ES cells, tumorigenesis.



Concepts (present)




Concepts (present)


(2) a. To exclude that Transitions (3) b. suggesting


Concepts (present)


(2) a. To exclude that Transitions (3) b. suggesting


Experiment (past)

Overlap with KEfEd

Facts

Concepts Problem
Implication
Hypothesis

Transitions Goal These results suggest that....

Method
Experiment
Results

More papers: citation erodes hedging

To investigate the possibility that
miR-372 and miR-373 suppress the
expression of LATS2, we...

Hypothesis


KnownFact KnownFact

Concepts Hypothesis


KnownFact KnownFact

Concepts Hypothesis

Goal

Method Result

Data

Experiment 1


Therefore, these results point to
LATS2 as a mediator of the miR-372 and
miR-373 effects on cell proliferation and
tumorigenicity,

KnownFact KnownFact

Concepts Hypothesis Implication

Goal

Method Result

Data

Experiment 1

Voorhoeve, 2006

tumorigenicity,

KnownFact KnownFact


Goal

Method Result

Data

Experiment 1

Voorhoeve, 2006

tumorigenicity,

KnownFact KnownFact


Goal
Goal

Method Result Method Result

Data Data

Experiment 1 Experiment 2

Voorhoeve, 2006
Raver-Shapira et.al, JMolCell 2007

Therefore, these results point to two miRNAs, miRNA-372 and-373, function as
LATS2 as a mediator of the miR-372 and potential novel oncogenes in testicular germ
miR-373 effects on cell proliferation and cell tumors by inhibition of LATS2 expression,
tumorigenicity, which suggests that Lats2 is an important
tumor suppressor (Voorhoeve et al., 2006).

KnownFact KnownFact

Concepts Hypothesis Implication Fact Fact

Goal
Goal


Data Data


More papers: citation erodes hedging Yabuta, JBioChem 2007
miR-372 and miR-373 target the
Voorhoeve, 2006 Lats2 tumor suppressor
(Voorhoeve et al., 2006)


KnownFact KnownFact


Goal
Goal


Data Data


More papers: citation erodes hedging Yabuta, JBioChem 2007
miR-372 and miR-373 target the
Voorhoeve, 2006 Lats2 tumor suppressor
(Voorhoeve et al., 2006)


KnownFact KnownFact


Goal
Goal


Data Data

http://elsatglabs.com/labs/anita/demos/LATSDemo102007/

“[Y]ou can transform a fact into ﬁction or a ﬁction into fact just by
adding or subtracting references [and data]”
– Bruno Latour, ‘Science in Action’,1987

Kinnings et al, 2010: Abstract hedged by editors
Authorʼs version Final version Changes Edit
1.!We have developed a computational 1.!We report a computational approach have developedʼ -> ʻreportʼ Hedge
approach that integrates structural that integrates structural ʻthe protein-drugʼ -> ʻa drug- Content
bioinformatics, molecular modeling bioinformatics, molecular modelling targetʼ
and systems biology to construct the and systems biology to construct a Proteome scale -> English
protein-drug network on a structural drug-target network on a structural proteome-wide scale
proteome scale. proteome-wide scale.
9. The Drugome-TB reveals not only 1.!The TB-drugome reveals that not only that aroundʼ -> English
that around one-third existing drugs approximately one-third of the drugs ʻapproximatelyʼ
show the potential to be repositioned to examined have the potential to be ʻof existing drugsʼ -> ʻof Hedge
treat tuberculosis, but also that many repositioned to treat tuberculosis and the drugs examinedʼ
currently unexploited M.tb proteins may that many currently unexploited M.tb ʻbut alsoʼ -> ʻandʼ English
be highly druggable and could receptors may be chemically ʻhighlyʼ -> ʻchemicallyʼ Hedge
therefore serve as novel antitubercular druggable and could serve as novel ʻthereforeʼ -> … English
targets. anti-tubercular targets.
6. It supports that the drug-target 6.#Indeed, our results support the It supports thatʼ -> ʻIndeed, Hedge/
network is inherently modular. idea that drug-target networks are our results support the idea English
inherently modular and further that that ʻ
any observed randomness is mainly
7.!The randomness observed in the caused by biased target coverage. …. -> ʻOur results support Hedge
network is mainly due to the false the idea that ʻ
drug-target connection and the biased ʻdue toʼ -> ʻmainly caused Hedge
target coverage. byʼ
10.With the advance of structural 8. More generally the methodology the methodology is ready for Hedge
genomics, the methodology is may be applied to other pathogens of other pathogen genomesʼ ->
ready for other pathogen genomes. interest with results improving as ʻMore generally the
more of their structural proteomes methodology may be
are determined through the applied... Content
continued efforts of structural ʻWith the advance of
biology/genomics. structural genomicsʼ -> ʻwith
results improving...
ʻmore ...are determined...ʼ
10 sentences 8 sentences Edits:14 Hedges
11 English corrections
4 Content corrections

Xie et al 2009: Claim gets validated in citations
Xie L, Li J, Xie L, Bourne PE, 2009:
Drug Discovery Using Chemical Systems Biology: Identification of the Protein-
Ligand Binding Network To Explain the Side Effects of CETP Inhibitors.
PLoS Comput Biol 5(5): e1000387. doi:10.1371/journal.pcbi.1000387

[Goal]: To understand the origins of these adverse drug reactions from Torcetrapib and other
related drugs undergoing clinical trials,
[Method]: we introduce a systematic strategy
[Goal] to identify off-targets in the human structural proteome and
[Goal]: investigate the roles of these off-targets in impacting human physiology and
pathology
[Method]: using biochemical pathway analysis.
[Reg-Imp]: Our findings suggest that
[Implication]: potential side-effects of a new drug can be identified at an early stage of
the development cycle and
[Implication]: [these can] be minimized by fine-tuning multiple off-target interactions.
[Reg-Hyp]: The hope is that
[Hypothesis]: this can reduce both the cost of drug development and the mortality rates
during clinical trials.

2010: 28 articles citing this paper; 19 full-text papers found.

2009: Paper enters knowledge space

1. Network analyses in systems pharmacology SI Berger… - Bioinformatics, 2009 - Oxford Univ Press
“Whereas originally medications were thought to hit a specific target and modulate effects through pathways
downstream of that target, we now know that many drugs hit multiple targets, each of which exist within a complex
network. The effects of the drug, both therapeutic actions and adverse events, are therefore a result of
perturbation of the complex network landscape. This was suggested in a study by Xie et al. (2009) where
structural homology between protein drug binding sites was used to predict potential off targets of Cholesteryl Ester
Transfer Protein Inhibitors.”

2. A unified statistical model to support local sequence order independent similarity searching for ligand-binding
sites and its application to genome-based drug … L Xie, L Xie… - Bioinformatics, 2009 - OUP
“Moreover, recently we have shown that such ligand–binding-site mapping can contribute to the design of
pharmaceuticals through the detection of off-targets (Kinnings et al., 2009; Xie et al., 2009; Xie et al., 2007).”
“Owing to scalability and robustness of SOIPPA and the EVD model, we have successfully applied SMAP to drug
discovery, including elucidation of the molecular mechanisms of drug side-effects and repurposing of safe
pharmaceuticals to target different pathways (Kinnings et al., 2009; Xie et al., 2009; Xie et al., 2007).“
“Using SMAP, we have successfully identified off-targets for several pharmaceuticals either already on the market or
in clinical trials. In one case we have revealed a complex off-target-binding network for cholesteryl ester transporter
protein (CETP) inhibitors (Xie et al., 2009). “

2. A unified statistical model to support local sequence order independent similarity searching for ligand-binding
sites and its application to genome-based drug … L Xie, L Xie… - Bioinformatics, 2009 - OUP
“Moreover, recently we have shown that such ligand–binding-site mapping can contribute to the design of
pharmaceuticals through the detection of off-targets (Kinnings et al., 2009; Xie et al., 2009; Xie et al., 2007).”
“Owing to scalability and robustness of SOIPPA and the EVD model, we have successfully applied SMAP to drug
discovery, including elucidation of the molecular mechanisms of drug side-effects and repurposing of safe
pharmaceuticals to target different pathways (Kinnings et al., 2009; Xie et al., 2009; Xie et al., 2007).“
“Using SMAP, we have successfully identified off-targets for several pharmaceuticals either already on the market or
in clinical trials. In one case we have revealed a complex off-target-binding network for cholesteryl ester transporter
protein (CETP) inhibitors (Xie et al., 2009). “
9. Predicting drug side-effects by chemical systems biology NP Tatonetti, T Liu… - Genome biology, 2009
“The recent work of Xie et al. [7] is another excellent example of the use of networks combining proteins and drugs.
They investigated the reasons for the serious side-effects of torcetrapib, an inhibitor of cholesteryl ester transfer protein
(CETP) that was in clinical trials as a preventive treatment for cardiovascular disease. The aim of torcetrapib was to raise
the levels of the desirable high-density lipoprotein cholesterol (HDL-C), but torcetrapib turned out to have the side-effect
of raising blood pressure, with potentially fatal effects in high-risk patients, and was withdrawn from development in 2006.
“Xie et al. [7] generated off-target binding networks by comparing the structure of ligand-binding sites in all known
protein structures. The proteins identified as having similar binding domains were ranked by a normalized docking score
and clustered by their structural and functional characteristics into a gene network that includes metabolic and regulation
pathways. Using this analysis, the authors identified possible off-targets for torcetrapib even though the binding
site of CETP itself is not fully described. Perhaps most interestingly, they incorporated biological pathways into their
off-target networks and found a potential explanation for the poorly understood effects of torcetrapib on blood pressure.
By combining a simple gene regulation model with the predicted binding affinities to activators and inhibitors of the renin-
angiotensin-aldosterone system (RAAS), they showed that torcetrapib caused more severe effects since it has a higher
affinity for more RAAS activators.”

2010: Paper is validated, contrasted and expanded

16. A Structure-Based Approach for Mapping Adverse Drug Reactions to the Perturbation of Underlying
Biological Pathways Wallach, N Jaitly, R Lilien… - PloS one, 2010 “Recently, Xie et al. [9] have used virtual docking to
study ADRs related to the cholesteryl ester transfer protein (CETP) and identified possible off-target interactions for a
set of CETP inhibitors. Their method uses a known 3D protein structure of the primary target to characterize the binding
site of the drugs. Then, it identifies potential off-targets by searching for other proteins having similar binding sites. Possible
interactions between the resulting set of proteins and the drugs are then studied using virtual docking.”
“In this work, we developed a computational framework for proposing associations between the ADRs of clinically approved
drugs and the modulation of underlying biological pathways. In contrast to the work of Xie et al. [9] and Yang et al. [12], our
model exploits the assumption that drugs capable of modulating similar pathways may have similar ADR profiles.”

19. The Mycobacterium tuberculosis Drugome and Its Polypharmacological Implications SL Kinnings, L Xie, KH
Fung, RM Jackson… - PLoS Computational …, 2010
“This strategy has been applied to study several selected drug targets, and proven, both computationally and
experimentally, to be a useful tool in drug repositioning [15], side effect prediction [16], [17], and polypharmacological target
discovery [18]. In this paper, we extend this methodology to the construction of a proteome-wide drug-target network. “


24. PROMISCUOUS: a database for network-based drug-repositioning J von Eichborn, MS Murgueitio… -
Nucleic Acids …, 2010 - Oxford Univ Press
As is the case with the aforementioned drug, until now most cases of drug repositioning are the result of serendipitous
observations. […] Another example was that of Xie et al. (16) in explaining the off-target effects of the CETP inhibitor
Torcetrapib that was taken out of phase III clinical trials after it was discovered to have significant side-effects.
Their work revealed a complex network of interactions with up to twelve putative off-targets.


24. PROMISCUOUS: a database for network-based drug-repositioning J von Eichborn, MS Murgueitio… -
Nucleic Acids …, 2010 - Oxford Univ Press
As is the case with the aforementioned drug, until now most cases of drug repositioning are the result of serendipitous
observations. […] Another example was that of Xie et al. (16) in explaining the off-target effects of the CETP inhibitor
Torcetrapib that was taken out of phase III clinical trials after it was discovered to have significant side-effects.
Their work revealed a complex network of interactions with up to twelve putative off-targets.

27. Cause-effect relationships in medicine: a protein network perspective AF Fliri, WT Loging… - Trends in
Pharmacological Sciences, 2010 – Elsevier
“Xie and coworkers used the similarity of cholesteryl ester transfer protein (CETP) ligand binding to identify off-target
proteins, and used these proteins to generate protein networks in an attempt to rationalize side-effect differences among
the CETP inhibitors torcetrapib, anacetrapib, and JTT-705 [89 L. Xie et al.], Drug discovery using chemical systems
biology: Identification of the protein-ligand binding network to explain the side effects of CETP inhibitors, The latter, the
most ‘promiscuous’ of the three, did not cause hypertension as a side effect whereas the more selective torcetrapib did.
This result suggested that the adverse events of medicines may not necessarily be reduced by increasing the target
selectivity of a drug but instead by ‘fine-tuning’ the pertinent cross-pathway communication. “

Hedging rules fact creation
- Small case study, but nonetheless clear:
- Authors need to downplay claims to be
acceptable in the literature
- After publication, hedging ‘erodes’, modality is
dropped: claim becomes fact.
- Danger: experimental context is not taken into
account when citing - claim gets extended
without experimental validation
- So how do we develop systems that allow
the context to stick to the claim?

A closer look at citations
!

Model of phenomenon
time = t+1

?

Model of phenomenon
time = t

Domain Model

(C) Ed Hovy

!

Model of phenomenon
time = t+1

?
Citation:
Model of phenomenon • ‘Lifts’ claim from previous paper
and add it to current paper
time = t • We need ‘rules for legal lifting’:
• Express author’s claim truly
Domain Model • Maintain (experimental) content
• Apply appropriately

(C) Ed Hovy

!

Model of phenomenon
time = t+1

?
Citation:
Model of phenomenon • ‘Lifts’ claim from previous paper
and add it to current paper
time = t • We need ‘rules for legal lifting’:
• Express author’s claim truly
Domain Model • Maintain (experimental) content
• Apply appropriately
What does this mean?
(C) Ed Hovy

The logical structure of a citation

(C) Ed Hovy


- The original author states
- a Proposition P : about entities, relations, events, actions…
- an Interpretation of P IA(P):
- Holder = author A
- Topic = P
- Context = conditions in area/domain for which P
- Epistemic valence = {true, hypothesized, untrue, ...}
- Strength/certainty/intensity/probability = ...
- Provenance: date, paper, journal ...

(C) Ed Hovy


- The original author states
- a Proposition P : about entities, relations, events, actions…
- an Interpretation of P IA(P):
- Holder = author A
- Topic = P
- Context = conditions in area/domain for which P
- Epistemic valence = {true, hypothesized, untrue, ...}
- Strength/certainty/intensity/probability = ...
- Provenance: date, paper, journal ...

- Citation: The citer’s interpretation of P IC(P) or of IA(P)
IC(IA(P)):
- Holder = citer C
- Topic = P or IA(P) ... etc.
(C) Ed Hovy

Rules for legal lifting

(C) Ed Hovy


- Citer must include Holder, Provenance, Topic,
Valence,

“De Waard (1862) showed/believed that P”

- Citer may include own interpretation IC(IA(P)) too:

(C) Ed Hovy


Valence,



“…but was [I believe] sadly mistaken”

(C) Ed Hovy


Valence,




- Citer may include Context, but must make certain
that it applies in the context described in the
current paper … This is often not done

(C) Ed Hovy


Valence,




- Citer should include the most canonical reference

(C) Ed Hovy


Valence,




- Citer really should include all relevant references

(C) Ed Hovy


Valence,

Citation
- Citer may include own interpretation IC(IA(P)) too: manage-
ment

tools can
- Citer may include Context, but must make certain assist
with all
that it applies in the context described in the of this
- Citer really should include all relevant references

(C) Ed Hovy

‘Referring Use of Tense’
Concepts
‘Gnomic’ present

Argumentation:
Instantaneous
present/To-inf
Other/Past Future
Work Discourse Work
Present progression Future
Instantaneous
present

Experiment
Event past

Anecdotal excursion into mythological text:

Anecdotal excursion into mythological text:
Facts in the Endogenous small RNAs (miRNAs) I sing of golden-throned Hera whom Rhea
eternal regulate gene expression by mechanisms bare. Queen of the immortals is she, surpassing
conserved across metazoans. all in beauty: she is the sister and the wife of
present: loud-thundering Zeus, --the glorious one whom
all the blessed throughout high Olympus
reverence and honor.
Events in the Vehicle-treated animals spent equivalent Now the wooers turned to the dance and to
simple past: time investigating a juvenile in the first and gladsome song, and made them merry, and
second sessions in experiments conducted in waited till evening should come; and as they
the NAC and the striatum: T1 values were made merry dark evening came upon them.
122 ± 6 s and 114 ± 5 s.
Events with We also generated BJ/ET cells expressing And she took her mighty spear, tipped with
embedded the RASV12-ERTAM chimera gene, which sharp bronze, heavy and huge and strong,
is only active when tamoxifen is added (De wherewith she vanquishes the ranks of men-of
facts: Vita et al, 2005). warriors, with whom she is wroth, she, the
daughter of the mighty sire.
Attribution in miRNAs have emerged as important In this book I have had old stories written
the present regulators of development and control down, as I have heard them told by intelligent
processes such as cell fate determination and people, concerning chiefs who have held
perfect: cell death (Abrahante et al., 2003, Brennecke dominion in the northern countries, and who
et al., 2003, Chang et al., 2004, Chen et al., spoke the Danish tongue; and also concerning
2004, Johnston and Hobert, 2003, Lee et al., some of their family branches, according to
1993, ... what has been told me.
Implications These results indicate that although Now it is said that ever since then whenever
are hedged, & miR-372&3 confer complete protection to the camel sees a place where ashes have been
oncogene-induced senescence in a manner scattered, he wants to get revenge with his
present tense similar to p53 inactivation, the cellular enemy the rat and stomps and rolls in the ashes
+ that: response to DNA damage remains intact. hoping to get the rat.

Trees of representation
Concepts
‘Present

Argumenta
tion:
present
Other/ Future
Past Discour Work
Present se Future
Present

Experiment
Past

Hedging, Fact from claim, Science as mythology

Recommended

Recommended

More Related Content

More from Anita de Waard

More from Anita de Waard (20)

Recently uploaded

Recently uploaded (20)

Hedging, Fact from claim, Science as mythology