# Journées annuelles du groupe de travail (2ème édition)

### Informations générales

Date : 1er et 2 juillet 2016

Lieu :
Salle "conférence",
1, place de l'École (en rez de chaussée, en dessous de la Maison
des mathématiques et de l'informatique, en face de l'amphithéâtre
Charles Mérieux),
69007 Lyon

Plan d'accès ici

Organisateurs : Olivier Gandrillon, Cédric Lhoussaine, Élisabeth Remy, Sylvain Sené et Anne Siegel

La deuxième édition des journées annuelles du GT Bioss va
se dérouler à la suite des Journées ouvertes de biologie,
informatique et mathématiques
(JOBIM)
organisées par la Société française de bio-informatique
(SFBI). Ainsi, les 1er et 2
juillet 2016, les membres du GT auront le plaisir de se rencontrer
autour de conférences autour des thèmes suivants :

- la modélisation stochastique en biologie ;

- la régulation génétique ;

- le métabolisme.

### Inscription

L'inscription, gratuite mais obligatoire, se fait en remplissant le formulaire accessible ici.

### Orateurs invités

Grégory BATT, INRIA Saclay

Marcelline KAUFMAN, Université libre de Bruxelles

Marie-France SAGOT, INRIA Lyon

### Programme

#### Vendredi 1er juillet

**09h00 - 09h30** - **Accueil**

**09h30 - 09h45** - **Introduction des journées**

**09h45 - 10h30** - **Conférence plénière** - Grégory Batt -
*Predicting long-term effects of apoptosis-inducing drug
treatments: coupling signal transduction pathways with
stochastic protein turnover models *

**10h30 - 11h00** - Bertrand Miannay -
*Identification des voies de signalisation impliquées dans
le myélome multiple par programmation par contrainte*

**11h00 - 11h30** - Arnaud Bonnaffoux -
*Toward a dynamic multi-scale/level approach for gene regulatory
network inference*

**11h30 - 12h00** - Nicolas Schabanel -
*Folding Turing is hard but feasible*

**12h00 - 13h30** - **Pause déjeuner**

**13h30 - 14h15** - **Conférence plénière** -
Marie-France Sagot -
*Species interactions from a metabolism perspective*

**14h15 - 14h45** - Nils Giordano -
*Dynamical allocation of cellular resources as an optimal
control problem*

**14h45 - 15h15** - Victorien Delannée -
*A modeling approach to evaluate the balance between
bioactivation and detoxification of MeIQx in human
hepatocytes*

**15h15 - 15h45** - **Discussion Bioss / GDR**

**15h45 - 16h15** - **Pause**

**16h15 - 16h45** - Hugues Berry -
*Estimating the effects of spatial non-homogeneities in
intracellular diffusion-reactions*

**16h45 - 17h15** - Dan Goreac -
*Hybrid designing using stochastic backward
equations*

**17h15 - 17h45** - Guillaume Madelaine -
*Structural simplifications of reaction networks: the
confluence problem*

**17h45 - 18h15** - Ferdinanda Camporesi -
*Context-sensitive flow analyses: a hierarchy of model
reductions*

#### Samedi 2 juillet

**09h00 - 09h45** - **Conférence plénière** -
Marcelline Kaufman -
*On multistationarity in chemical reaction networks*

**09h45 - 10h15** - Kévin Perrot -
*On the flora of asynchronous locally non-monotonic
Boolean networks*

**10h15 - 10h45** - Élisabeth Remy -
*Discrete dynamics of compound regulatory circuits*

**10h45 - 11h00** - **Pause**

**11h00 - 11h30** - Loïc Paulevé -
*Around reachability in automata networks*

**11h30 - 12h00** - Emna Ben Abdallah -
*Inference of biological regulatory networks from time
series data*

**12h00 - 12h30** - Adrien Richard -
*Points fixes dans les réseaux booléens monotones*

### Résumés

**Grégory Batt** - *Predicting long-term effects of
apoptosis-inducing drug treatments: coupling signal
transduction pathways with stochastic protein turnover
models*

TRAIL is an anti-cancer drug that induces apoptosis
selectively in cancer cells. Unfortunately even high doses
of TRAIL do not kill all cells and subsequent TRAIL
treatments are transiently less effective. Despite extensive
studies, a mechanistic understanding of these phenomena is
still lacking. In this talk, I will present an extension of
a previously-proposed model describing TRAIL signal
transduction in Hela cells (Spencer et al, Nature 2011) with
models accounting for the turnover of the proteins involved
in the pathway at the cell level and the dynamics (growth
and death) of the cell population in monolayers or in 3D
spheroids. This model is minimalistic in the sense that it
uses default values from the literature for all but two
parameters. Yet, it explains the existence of survivors
(fractional killing), the increased resistance of the
surviving population and its transient aspect. The analysis
of model predictions calls into question the importance of
survival pathways and highlights the critical role of the
stochastic turnover of proteins in zymogen-based pathways in
which activated forms are rapidly degraded.

**Emna Ben Abdallah** - *Inference of biological
regulatory networks from time series data*

With the development of high-throughput data, there is a
growing need for methods that automatically generate
(resp. revise) admissible models. Our research aims at
providing a logical approach to infer Biological Regulatory
Networks based on given time series data. We propose a new
methodology for models expressed through a timed extension
of the Process Hitting framework (well suited for biological
systems). The main purpose is to have as a result the most
consistent network as possible with the observed data. The
originality of our work relies on the integration of
quantitative time delays directly in our learning
approach.

Taking as input a background knowledge under the form of
influence graph and time series data, the contribution of
our method lies in the fact that we identify the set of
actions between biological components by concretizing the
signs (negative or positive) besides providing thresholds
and associating the quantitative time delays. Starting from
the structure of the system and its experimental time
series, the method addresses both inference and revision:
(1) If no previous dynamic model is given, we infer the
dynamics of the system. (2) Otherwise we take profit from
new time series to revise actions and delays.

We will show the benefits of such automatic approach on
dynamical biological model, the circadian clock, and we
conduct benchmarks on the DREAM4 datasets, a popular
reverse-engineering challenge, in order to discuss the
computational performances of our algorithm.

**Hugues Berry** - *Estimating the effects of spatial
non-homogeneities in intracellular
diffusion-reactions*

The inner of living cells exhibits disorder, non-homogeneity
and obstruction. For instance, cell membranes are
heterogeneous collections of hierarchical spatial domains
with various length scales and timescales (e.g., fences,
lipid rafts, and caveolae) that spatially modulate the
diffusion of proteins. This defines a spatially
nonhomogeneous diffusion problem with position-dependent
diffusion coefficient. The impact of these deviations from
simple Brownian motion on the biochemical reactions that
take place in cells cannot be studied with the classical
mass-action laws of biochemical kinetics and are just
starting to be explored by spatially-explicit stochastic
simulations. In this talk, I will present an overview of the
recent modelling work carried out in our group on the
effects of receptor clustering on the dynamics of
ligand-binding equilibrium, and on correlations in gene
positions for repressilator-like gene regulation loops. Our
results suggest that spatial non-homogeneities are potent
modulators of the apparent affinity of the equilibrium
reaction or of the dynamical regime itself, even when the
elementary reaction rates are not altered.

**Arnaud Bonnaffoux** - *Toward a dynamic
multi-scale/level approach for gene regulatory network
inference*

Gene regulatory networks (GRN) play an important role in
many biological processes, such as differentiation, and
their identification has raised great expectations for
understanding cell behavior. Many computational GRN
inference approaches have been described, which are based on
expression data but they face common issues such as data
scarcity, high dimensionality or population blurring (Chai
et al., 2014). We believe that recent high-throughput single
cell expression data (see e.g. Pina et al., 2012 ; Shalek et
al., 2014) acquired in time-series will allow to overcome
these issues and give access to causality, instead of
« simple » correlation, for gene interactions. Causality is
very important for mechanistic model inference and
biological relevance because it enables the emergence of
cellular decision-making. Emergent properties of a
mechanistic model of a GRN should then match with
multi-scale (molecular/cellular) and multi-level (single
cell/population) observations. We will expose a GRN
inference framework based on these assumptions. It follows
three steps:

1. Node parametric inference. We have inferred
the parameters from a stochastic mechanistic model of gene
expression, the Random Telegraph model (Kim and Marioni,
2013), thank's to time-series single cell expression data
from a population of chicken erythrocyte progenitor during
their differentiation process (Gandrillon et al., 1999)

2. Model reduction. This is mostly an ongoing work, and will
make use of specific constraints applying to the network.

3. The final step will consist in network inference
constrained by dynamic multi-scale/level
observations.

**Ferdinanda Camporesi** - *Context-sensitive flow
analyses: a hierarchy of model reductions*

Rule-based modelling allows very compact descriptions of
protein-protein interaction networks. However, combinatorial
complexity increases again when one attempts to describe
formally the behaviour of the networks, which motivates the
use of abstractions to make these models more
coarse-grained. Context-insensitive abstractions of the
intrinsic flow of information among the sites of chemical
complexes through the rules have been proposed to infer
sound coarse-graining, providing an efficient way to find
macro-variables and the corresponding reduced models. In
this paper, we propose a framework to allow the tuning of
the context-sensitivity of the information flow analyses and
show how these finer analyses can be used to find fewer
macro-variables and smaller reduced differential
models.

**Victorien Delannée** - *A modeling approach to
evaluate the balance between bioactivation and
detoxification of MeIQx in human hepatocytes*

Heterocyclic aromatic amines (HAA) are environmental and
food contaminants that are potentially carcinogen for
human. 2-Amino-3-methylimidazo(4,5-f)-quinoxaline (MeIQx) is
one of the most abundant HAA formed in cooked meat. MeIQx is
metabolized by cytochrome P450 1A2 in human liver into
detoxification and bioactivation products. Once
bioactivated, MeIQx metabolites can lead to DNA adduct
formation responsible for further genome instability. Using
a computational approach, we develop a numerical model for
MeIQx metabolism that predicts the MeIQx biotransformation
into detoxification or bioactivation pathways according to
the concentration of MeIQx. Our model permits to
investigate the balance between bioactivation (i.e. DNA
adduct formation pathway through Ester-O-NH-MeIQx) and
detoxification of MeIQx in order to predict the behaviour of
this environmental contaminant in human liver.

Our results demonstrate that 1) the detoxification pathway
predominates, 2) predicting the bioactivation and the
detoxification for any initial concentration of MeIQx at any
time is feasible for any dataset and 3) the ratio between
detoxification and bioactivation pathways is not linear and
shows a maximum at 10µM of MeIQx in hepatocyte cell
model.

**Nils Giordano** - *Dynamical allocation of cellular
resources as an optimal control problem: novel insights
into microbial growth strategies*

Microbial physiology exhibits growth laws that relate the
macromolecular composition of the cell to the growth
rate. Recent work has shown that these empirical
regularities can be derived from coarse-grained models of
resource allocation. While these studies focus on
steady-state growth, such conditions are rarely found in
natural habitats, where microorganisms are continually
challenged by environmental fluctuations. The aim of this
paper is to extend the study of microbial growth strategies
to dynamical environments, using a self-replicator model. We
formulate dynamical growth maximization as an optimal
control problem that can be solved using Pontryagin’s
Maximum Principle. We compare this theoretical gold standard
with different possible implementations of growth control in
bacterial cells. We find that simple control strategies
enabling growth-rate maximization at steady state are
suboptimal for transitions from one growth regime to
another, for example when shifting bacterial cells to a
medium supporting a higher growth rate. A near-optimal
control strategy in dynamical conditions is shown to require
information on several, rather than a single physiological
variable. Interestingly, this strategy has structural
analogies with the regulation of ribosomal protein synthesis
by ppGpp in the enterobacterium Escherichia coli. It
involves sensing a mismatch between precursor and ribosome
concentrations, as well as the adjustment of ribosome
synthesis in a switch-like manner. Our results show how the
capability of regulatory systems to integrate information
about several physiological variables is critical for
optimizing growth in a changing environment.

**Dan Goreac** - *Hybrid designing using stochastic
backward equations*

We present some targeted-behaviour based issues in the
hybrid modelling of networks. The common method is derived
from the theory of BSDEs (backward stochastic differential
equations) by interpreting the reaction speeds as externally
controlled (thus, modifiable) parameters. In the case of
first-order (linear) models, we give explicit (algebraic)
conditions on the sets of parameters leading to
"controllability" (targeted behaviour). For more general
systems, if the time allows it, we give an intuition on how
parameters might be chosen by using reflected backward
equations and embedding in spaces of measures.

**Marcelline Kaufman** - *On multistationarity in
chemical reaction networks*

Résumé au format
pdf ici.

**Guillaume Madelaine** - *Structural simplifications
of reaction networks: the confluence problem*

Models in system biology are often big, and need to be
simplified in order to be analyzed, simulated or
verified. We will first present a set of simplification
rules for reaction networks without kinetic rates. This
simplification preserves the non-deterministic semantics, in
terms of reachability of final strongly connected
components, called attractors. Then we will extend the
reaction networks with kinetic rates. We will show that,
under partial steady-state assumptions, we can simplify the
networks by removing some linear intermediate molecular
species, while preserving the deterministic semantics of the
other species. We will focus on the confluence of this
simplification, that is do we always obtain the same fully
simplified network, independently of the order in which the
simplification rules are applied. We will show that removing
the linear intermediate species is not confluent in
general. By adding another rule that simplifies some
"dependent reactions", we will show that we can always
obtain the same structure of the network and the same
ODEs. However, the distribution of the kinetic rates between
the reactions can be different.

**Bertrand Miannay** - *Identification des voies de
signalisation impliquées dans le myélome multiple par
programmation par contrainte*

Résumé au format
pdf ici.

**Loïc Paulevé** - *Around reachability in automata
networks*

Many elaborated questions in systems biology involve the one
of reachability : the existence / inevitability of a
sequence of events leading from a state to another. Some
involve the verification of reachability, many more the
inference of mutations for its control. Reachability is a
difficult computational problem: it is PSPACE-complete for
Automata Networks / Petri nets with finite discrete state
space. Methods relying on network topology, concurrency,
abstract interpretation, model reduction, aim at coping with
reachability in large scale networks. In this talk, I'll
give an overview of a range of these methods and related
tools, with their applications to model-checking, dynamical
bifurcation identification, control target prediction, and
cellular differentiation.

**Kévin Perrot** - *On the flora of asynchronous
locally non-monotonic Boolean networks*

Studies on the dynamics of Boolean networks (BNs) have
mainly focused on monotonic networks, where fundamental
questions on the links relating their static and dynamical
properties have been raised and addressed. In this
presentation, we will explore analogous questions on
non-monotonic networks, xor-BNs, that are BNs where all the
local transition functions are xor-functions. Using
algorithmic tools, we will present a general
characterisation of the asynchronous dynamics for most of
the cactus xor-BNs and strongly connected xor-BNs, through
new bisimulation equivalences specific to xor-BANs.

**Élisabeth Remy** - *Discrete dynamics of compound
regulatory circuits*

In biological regulatory networks represented in terms of
signed, directed graphs, topological motifs such as circuits
are known to play key dynamical roles. We present results on
the dynamical impact of the addition of a short-cut in a
regulatory circuit. More precisely, based on a Boolean
formalisation of regulatory graphs, we have unrolled
complete descriptions of the discrete dynamics of particular
motifs, under the synchronous and asynchronous updating
schemes. These motifs are made of a circuit of arbitrary
length, combining positive and negative interactions in any
sequence, encompassing a short circuit, and using AND, OR
and XOR logical rules.

**Adrien Richard** - *Points fixes dans les réseaux
booléens monotones*

Les réseaux booléens sont des systèmes dynamiques où chaque
variable ne peut prendre que deux états possibles: 0 ou
1. Depuis les travaux pionniers de Kauffman et Thomas, ce
sont des modèles très classiques pour les réseaux de
gènes. Dans ce contexte, les points fixes sont d'un intérêt
particulier: ils correspondent à des patterns stables
d'expression des gènes souvent reliés à des fonctions
cellulaires bien précises. Cependant, les premières
informations disponibles sur un réseau de gènes concernent
généralement le graphe d'interaction du réseau et non sa
dynamique.

Une question naturelle est donc la suivante:
que peut-on dire sur les points fixes d'un réseau booléen en
fonction de son graphe d'interaction seulement ?

Dans cette exposé, on présente une étude du plus grand
nombre de points fixes qu'un réseau booléen monotone peut
admettre en fonction de son graphe d'interaction. On donnera
des bornes inférieures et supérieures qui ne dépendent que
de la structure des cycles du graphe d'interaction. Les deux
paramètres centraux seront, d'une part, la taille d'un plus
petit ensemble de sommets intersectant tous les cycles et,
d'autre part, le plus grand nombre de cycles
disjoints. L'étude fera intervenir des théorèmes, classiques
en combinatoire, sur le treillis booléen et ses antichaines.

C'est un travail réalisé en collaboration avec Julio Aracena
et Lilian Salinas (Université de Concepcion, Chili)
disponible à l'adresse suivante:
http://arxiv.org/abs/1602.03109.

**Marie-France Sagot** - *Species interactions from a
metabolism perspective*

The frontier between different species may be considered
very fuzzy as is more and more observed. Organisms are no
longer perceived as single genetically identical individuals
and are rather considered as part of communities. At its
extreme, one could see thus the whole of life as forming one
single community, or a community of communities interacting
sometimes closely and for long periods of evolutionary
time. Such interactions appear essential to understand some
if not all of the most fundamental evolutionary and
functional questions related to living organisms. They
however remain very little explored by computational
biologists, perhaps due to the difficult modelling and
computational issues raised. Yet, because of the complexity
and singularity of these communities, it is clear that
experimental data alone do not allow to fully understand the
global capacities and functions of these organisms and their
interactions. In this talk, I will briefly present some of
the models and algorithms, in the case related to
metabolism, that we have recently been developing with the
goal of better understanding some such close and often
persistent interactions. I will also mention a much longer
term objective of this work that would be to become able in
some cases to suggest the means of controlling for
equilibrium in an interacting community.

**Nicolas Schabanel** - *Folding Turing is hard but
feasible*

We introduce and study the computational power of Oritatami,
a theoretical model to explore greedy molecular folding, by
which the molecule begins to fold before waiting the end of
its production. This model is inspired by our recent
experimental work demonstrating the construction of shapes
at the nanoscale by folding an RNA molecule during its
transcription from an engineered sequence of synthetic
DNA. While predicting the most likely conformation is known
to be NP-complete in other models, Oritatami sequences fold
optimally in linear time. Although our model uses only a
small subset of the mechanisms known to be involved in
molecular folding, we show that it is capable of efficient
universal computation, implying that any extension of this
model will have this property as well.

We develop several general design techniques for programming
these molecules. Our main result in this direction is an
algorithm in time linear in the sequence length, that finds
a rule for folding the sequence deterministically into a
prescribed set of shapes depending of its environment. This
shows the corresponding problem is fixed-parameter tractable
although we proved it is NP-complete in the number of
possible environments. This algorithm was used effectively
to design several key steps of our constructions.

### Participants

Jalouli ACHREF, Université de Limoges

Vicente ACUNA, CMM, Santiago, Chili

Émilie ALLART, CRISTAL, Université de Lille

Adel Amar AMOURI, Dpt. de biologie, Université d'Oran

Emna BEN ABDALLAH, IRCCyN, École centrale de Nantes

Adrien BASSO-BLANDIN, LIP, ENS-Lyon

Grégory BATT, Lifeware, INRIA Saclay

Guillaume BEAUMONT, IPS2, Université Paris Sud

Emmanuelle BECKER, IRSET, Université de Rennes

Hugues BERRY, Beagle, INRIA Lyon

Arnaud BONNAFFOUX, LBMC, ENS-Lyon

Ferdinanda CAMPORESI, DIENS, ENS

Thomas COKELAER, Biomics, Institut Pasteur

Victorien DELANNÉE, IRISA, Université de Rennes

Ronan DUCHESNE, LBMC, ENS-Lyon

Maxime FOLSCHETTE, I3S, Université de Nice - Sophia Antipolis

Enrico FORMENTI, I3S, Université de Nice - Sophia Antipolis

Olivier GANDRILLON, LBMC, CNRS Lyon

Nils GIORDANO, INRIA Grenoble

Dan GOREAC, LAMA, Université Paris-Est Marne-la-Vallée

Carito GUZIOLOWSKI, IRCCyN, École centrale de Nantes

Pierre GUILLON, I2M, CNRS Marseille

Russ HARMER, LIP, ENS-Lyon

Ulysse HERBACH, LBMC, ENS-Lyon

Marcelline KAUFMAN, Dpt. de chimie physique et biologie théorique,
Université libre de Bruxelles

Cédric LHOUSSAINE, CRISTAL, Université de Lille

Guillaume MADELAINE, CRISTAL, Université de Lille

Bertrand MIANNAY, IRCCyN, École centrale de Nantes

Jean-Michel MULLER, LIP, CNRS Lyon

Loïc PAULEVÉ, LRI, CNRS Orsay

Kévin PERROT, LIF, Université d'Aix-Marseille

Arnaud PORET, LBMC, ÉNS-Lyon

Sylvain PRIGENT, Sysbio, Université de Chalmers

Élisabeth REMY, I2M, CNRS Marseille

Adrien RICHARD, I3S, CNRS Nice - Sophia Antipolis

Marie-France SAGOT, ERABLE, INRIA Lyon

Nicolas SCHABANEL, IRIF, CNRS Paris

Sylvain SENÉ, LIF, Université d'Aix-Marseille

Anne SIEGEL, IRISA, CNRS Rennes

Laurent TRILLING, TIMC-IMAG, Université de Grenoble

Jean-Yves TROSSET, BIRL, SupBioTech