RESEARCH Open Access
A memb ane compu ing simula o o
ans-hie a chical an ibio ic esis ance
e olu ion dynamics in nes ed ecological
compa men s (ARES)
Ma celino Campos
1,3
, Ca los Llo ens
2*
, José M. Sempe e
3
, Rica do Fu ami
2
, I ene Rod iguez
1,4,5
,
Pu i icación Ca asco
6
, Ra ael Capilla
2
, Ampa o La o e
5,6,7
, Te esa M. Coque
1,4,5
, And es Moya
5,6,7
and Fe nando Baque o
1,4,5*
Abs ac
Backg ound: An ibio ic esis ance is a majo biomedical p oblem upon which public heal h sys ems demand
solu ions o cons ue he dynamics and epidemiological isk o esis an bac e ia in an h opogenically-al e ed
en i onmen s. The implemen a ion o compu able models wi h ecip oci y wi hin and be ween le els o biological
o ganiza ion (i.e. essen ial nes ing) is cen al o s udying an ibio ic esis ances. An ibio ic esis ance is no jus he
esul o an ibio ic-d i en selec ion bu mo e p ope ly he consequence o a complex hie a chy o p ocesses
shaping he ecology and e olu ion o he dis inc subcellula , cellula and sup a-cellula ehicles in ol ed in he
dissemina ion o esis ance genes. Such a complex backg ound mo i a ed us o explo e he P-sys em s anda ds o
memb ane compu ing an inno a i e na u al compu ing o malism ha abs ac s he no ion o mo emen ac oss
memb anes o simula e an ibio ic esis ance e olu ion p ocesses ac oss nes ed le els o mic o- and mac o-
en i onmen al o ganiza ion in a gi en ecosys em.
Resul s: In his a icle, we in oduce ARES (An ibio ic Resis ance E olu ion Simula o ) a so wa e de ice ha
simula es P-sys em model scena ios wi h i e ypes o nes ed compu ing memb anes o ien ed o emula e a
hie a chy o eco-biological compa men s, i.e. a) pe iphe al ecosys em; b) local en i onmen ; c) ese oi o supplies;
d) animal hos ; and e) hos ’s associa ed bac e ial o ganisms (mic obiome). Compu a ional objec s emula ing
molecula en i ies such as plasmids, an ibio ic esis ance genes, an imic obials, and/o o he subs ances can be
in oduced in o his amewo k and may in e ac and e ol e oge he wi h he memb anes, acco ding o a se o
p e-es ablished ules and speci ica ions. ARES has been implemen ed as an online se e and o e s addi ional ools
o s o age and model edi ing and downs eam analysis.
(Con inued on nex page)
* Co espondence: [email p o ec ed];[email p o ec ed]
2
Bio ech ana, Valencia, CEEI Building, Benjamin F anklin A . 12, Valencia
Technological Pa k, 46980 Pa e na, Spain
1
Depa men o Mic obiology, Ramón y Cajal Uni e si y Hospi al, IRYCIS,
Ca e e a de Colmena Viejo, km. 9,100, 28034 Mad id, Spain
Full lis o au ho in o ma ion is a ailable a he end o he a icle
© 2015 Campos e al.
Open Access
This a icle is dis ibu ed unde he e ms o he C ea i e Commons A ibu ion 4.0
In e na ional License (h p://c ea i ecommons.o g/licenses/by/4.0), which pe mi s un es ic ed use, dis ibu ion, and
ep oduc ion in any medium, p o ided you gi e app op ia e c edi o he o iginal au ho (s) and he sou ce, p o ide a link o
he C ea i e Commons license, and indica e i changes we e made. The C ea i e Commons Public Domain Dedica ion wai e
(h p://c ea i ecommons.o g/publicdomain/ze o/1.0/) applies o he da a made a ailable in his a icle, unless o he wise s a ed.
Campos e al. Biology Di ec (2015) 10:41
DOI 10.1186/s13062-015-0070-9
(Con inued om p e ious page)
Conclusions: The s ochas ic na u e o he P-sys em model implemen ed in ARES explici ly links wi hin and
be ween hos dynamics in o a simula ion, wi h eedback ecip oci y among he di e en uni s o selec ion
in luenced by an ibio ic exposu e a a ious ecological le els. ARES o e s he possibili y o modeling p edic i e
mul ile el scena ios o an ibio ic esis ance e olu ion ha can be in e oga ed, edi ed and e-simula ed i necessa y,
wi h di e en pa ame e s, un il a co ec model desc ip ion o he p ocess in he eal wo ld is con incingly
app oached. ARES can be accessed a h p://gydb.o g/a es.
Re iewe s: This a icle was e iewed by Eugene V. Koonin, and E ic Bap es e.
Keywo ds: Memb ane compu ing, P-sys em, An ibio ic esis ance, Essen ial nes ing
Backg ound
An ibio ic esis ance (AR) is a se ious biomedical p oblem
upon which public heal h sys ems u ge o new R&D
s a egies o p e en he eme gence and dissemina ion o
esis an bac e ia in human inhabi ed en i onmen s [1–6].
The disco e y and clinical use o an imic obials du ing
he pas 20 h cen u y has changed he cou se o medicine
and he human li es yle by educing mo ali y om bac-
e ial in ec ions. Howe e , he use o misuse o an ibio ics,
medicines and d ugs may lead bac e ia o become ole an
o he ac ion o an imic obials, e en ually making s and-
a d ea men s ine ec i e and hus challenging key med-
ical p ac ices as in ensi e ca e medicine, ansplan a ion,
o he he apy o immuno-comp omised pa ien s [2, 5, 7].
AR is also a subjec o pa icula in e es o ood-chain
s akeholde s who use an imic obials as g ow h-p omo ing
addi i es un il ecen ly o o p e en ing c oss-in ec ion in
ood p oduc ion a eas [7–12]. Unde an ibio ic exposu e,
a ms ha e become signi ican ese oi s o an ibio ic e-
sis an mic oo ganisms, which in u n can be ansmi ed
om ood-p oducing animals and plan s o humans [7–13].
The ma e is o unde s and he isks and he unin ended
consequences o he an h opogenic use and elease o
an imic obial agen s in o he biosphe e, which wi hou ap-
p op ia e measu es migh d i e he wo ld owa ds a pos -
an ibio ic e a whe e mo ali y due o un ea able and a al
in ec ions may become cus oma y, pa icula ly in unde -
de eloped a eas o he wo ld [3–5, 12, 14]. Unde s anding
AR e olu ion is howe e a complex issue, since i is no
jus he esul o an ibio ic-d i en selec ion o mu an e-
sis an bac e ial clones as i is equen ly conside ed, bu
mo e p ope ly he consequence o a a ie y o ans-
hie a chical in e ac ions be ween all biological ehicles
in ol ed in he dissemina ion o he gene ic in o ma ion
in ol ing AR [15–18]. We a e alking abou he ollowing
biological en i ies: gene ic pla o ms, ansposons and/o
plasmids (he e called subcellula eplica o s), bac e ial
cells, gene ic exchange communi ies (GEC, communi ies
in which he in e change o gene ic ma e ial occu s e-
quen ly), mic obio a, hos indi iduals and hos communi-
ies. In o he wo ds, no only AR genes can be ca ie s o
AR gene ic in o ma ion, bu also all o he biological uni s
in which AR genes can be successi ely loca ed a di e en
subcellula , cellula and sup a-cellula nes ed le els o he
ecosys em [19–21]. I is impo an o no e he need o un-
de s anding he e ec o he whole nes ed ames associ-
a ed o AR o es ablish a comp ehensi e “pa ame e
space”able o desc ibe he mul idimensional e olu ion o
an ibio ic esis ance (Table 1, o mo e de ails see [22]).
All he a o esaid ca ie s a e uni s o selec ion ha can be
simul aneously and independen ly chosen a di e en en-
i onmen al le els including “in i onmen al”o mic obio-
ic ecosys ems [23]. The e o e, we should expec a high
complexi y in be ween-hos s demog aphical dynamics, as
he hos colonized o in ec ed by esis an bac e ia usually
Table 1 Key-nes ed ames associa ed o AR: a complex
pa ame e space
a) Densi y o colonized and colonizable hos s wi h an ibio ic esis an
bac e ia
b) Popula ion sizes o bac e ia pe hos du ing coloniza ion and in ec ion
c) Suscep ibili y o coloniza ion o hos s, including age, gende ,
e hnici y, nu i ion,illness- acili a ed coloniza ion
d) F equency o be ween-hos s in e ac ions i.e. ,human- o-human o
animal-human in e ac ions
e) Hos na u al and acqui ed immune esponse o colonizing o ganisms
) Ecological pa ame e s o colonizable a eas, including in e ac ion wi h
local mic obio a and equency and ype o an ibio ic- esis an
commensals
g) Mig a ion and dispe sal
h) An ibio ic and biocide exposu e and o e all densi y o an ibio ic use,
ype o an ibio ics and mode o ac ion, dosage and du a ion o
he apy, adhe ence o he apy, selec i e an ibio ic concen a ions,
an ibio ic combina ions
i) Mode o ansmission o esis an o ganisms om he en i onmen o
hos s
j) T ansmission a es be ween hos s (an ibio ic ea ed and no - ea ed,
in ec ed, and no -in ec ed)
k) Time o con ac be ween hos s
l) Hygiene, in ec ion con ol, sani a ion
m) Food, and d inking wa e con amina ion by esis an bac e ia and
hos exposu e
n) En i onmen al con amina ion by esis an o ganisms, including soil,
sewage and wa e
Campos e al. Biology Di ec (2015) 10:41 Page 2 o 13
belongs o a popula ion o in e ac ing indi iduals ha in
u n belong o a communi y o in e ac ing bac e ial spe-
cies. No e, o ins ance, ha lice and keds o pe s and a m
animals (Melophagus o inus, Linogna hus i uli, He e o-
doxus spinige ) o pes -insec s such as he house ly
(Musca domes ica) and he Ge man cock oach Bla ella
ge manica, which a e known o play a signi ican ole as
ese oi s and ec o s o oppo unis ic bac e ial pa hogens
ha a e o en esis an o an ibio ics [24–26] as hey
mo e, eely and indisc imina ely, om il h and animal
was e o ood, allowing esis an bac e ia o explo e new
habi a s in hospi als, communi y se ings and ood acil-
i ies [27–29]. This ype o complex scena ios, whe e bac-
e ial esis an popula ions and hei gene ic pla o ms
con aining AR genes mo e because hey a e selec ed, and
e ol e because o hei p omiscuous mig a ion, es ablish
he p econdi ion o isualize AR no only as a ma e o
biological unc ion, bu also as in o ma ion low p ocess-
ing. Fo hese easons, AR e olu ion is wi hou doub one
o he majo biomedical challenges o esea che s in epi-
demiology and sys ems biology.
In e es ingly, om he epis emological in e ac ion o
sys em biology, compu e science and ma hema ics, a
a ie y o models ha e a isen in he las decades o con-
nec pe o mances a di e en scales. These models a e
known as nes ed o embedded models ( o a e iew see
[30]) and ha e been used o accep ably add ess speci ic
ques ions in ol ing wi hin-hos dynamics enclosed in a
model o be ween-hos epidemiological scena ios. Nes ed
models a e classi ied as “inessen ial”when he wi hin-hos
dynamic in luences be ween-hos p ocesses bu no ice
e sa,o “essen ial”when he e is a ecip ocal eedback
be ween le els o o ganiza ion. In pa icula , AR modeling
equi es an essen ial nes ed model; any al e a ion o he
ca ie s in any speci ic esis ance ai , o in hei mecha-
nisms o a ia ion and mobiliza ion (mu a ion, ecombin-
a ion, ansposi ion, ho izon al gene ans e , mig a ion)
may in luence he dynamics o o he uni s o highe and
lowe hie a chy, ha ing logical consequences on he e-
quency and dissemina ion o AR genes and, he e o e,
e olu iona y and/o ecological consequences on bac e ial
popula ion [31, 32]. Un il no oo long ago, he di icul y
o model his ype o scena ios wi h essen ial nes ing was
an impo an limi a ion o easibly s udy AR e olu ion
p ocesses. Howe e , exci ing new oppo uni ies ha e e-
cen ly a isen om a na u al compu ing o malism inspi ed
on he s uc u e and unc ioning o biological cells, called
memb ane compu ing [33–35]. Memb ane compu ing
concei es any biological sys em as a hie a chical cons uc
whe e he low o ma e ials can be in e p e ed as compu -
ing p ocesses. In pa icula , memb ane compu ing o e s a
e sa ile amewo k known as P-sys em ha consis s o a
hie a chical memb ane s uc u e o nes ed compa men s
whe e mul ise s o objec s a e loca ed and can mo e
ac oss memb anes e ol ing acco ding o a ini e numbe
o gi en ules. Memb ane compu ing ha e been p o ed o
be uni e sal models o compu a ion [34] and has been
success ully used o model oscilla o y sys ems [36], p o-
cesses o signal ansduc ion [37, 38], gene egula ion con-
ol [39], quo um sensing [40], me a-popula ions [41] and
ecosys ems [42, 43] hus sugges ing ha any hing ha can
be compu ed can be done so as a P-sys em. Fo mo e de-
ails on he di e en app oaches eached unde memb ane
compu ing, see [44], o e e o he o icial websi es o he
memb ane compu ing communi y [45] and P–Lingua [46]
he p og amming language used o he de elopmen o
P-sys ems.
In his pape we in oduce a new P-sys em model de-
signed o compu ing a h ee le els o o ganiza ion
(subcellula , cellula , sup a-cellula ) h ough he so wa e
implemen a ion o a simula o we call An ibio ic Resis -
ance E olu ion Simula o (ARES). The gene al aim o
ARES is o acili a e p edic i e compu a ional models on
he po en ial ans-hie a chical esponse o AR o pa -
icula in e en ions in speci ic scena ios. The simula-
o ´s p ojec is a wo k in p og ess, equi ing cons an
e inemen s de i ed om he expe iences (“expe i-
men s”) o cos ume s. The i s e sion he e in oduced,
is a p o o ype ha o e s a p ede ined layou composed
o i e ypes o nes ed-memb anes ha concep ually
emula e an ecosys em hie a chy o biological bounda ies
based on popula ion en i onmen al a eas, ese oi s,
hos popula ions and bac e ial lineages o oppo unis ic
pa hogens. G an ed o he implemen a ion a iendly- o-
use on -end in e ace, he use is allowed o de ine a
s a ing con igu a ion o elemen s (subcellula ehicles,
an imic obials and o he subs ances) inhabi ing he
a o esaid memb anes, speci ica ions and ules acco ding
o which bo h elemen s and memb anes e ol e h ough
a numbe o i e a ions. ARES is hos ed a he GyDB
P ojec [47] a da abase o esea ch o mobile gene ic
elemen s ( ele an ca ie s in he s udy o AR), and
has been launched as an online se e accessible a
h p://gydb.o g/a es.
Me hods
P-sys em model o simula ing ecosys ems wi h nes ed
ecological bounda ies
AR is a p ocess o mul ile el selec ion o nes ed uni s
whe e he dis inc esis ance-ca ie s (gene ehicles) in-
luence each o he o selec ion and in og essi e c oss-
ing o esis ance o an ibio ics a di e en en i onmen al
le els (subcellula , cellula and sup acellula ) [18, 20, 48].
This is an eco-biological model ha can be o mally gen-
e alized acco ding o he ollowing uple:
Y¼V;μ;w1;w2;…;wn;R1;ρ1
ðÞ;…Rn;ρn

Campos e al. Biology Di ec (2015) 10:41 Page 3 o 13
whe e Vis a wo king alphabe o objec s; μis a memb ane
s uc u e consis ing o nmemb anes labeled 1, 2, . . . , n
ep esen s a oo ed ee; and w
1
,...,w
n
a e s ings o e
V ha ep esen mul ise s o objec s ini ially placed in he
s uc u e o nmemb anes, which, om ha poin on will
be e e ed as ecological bounda ies (EBs) h ough he es
o his a icle.
In he Valphabe , ou model akes in o accoun he
ollowing en i ies o be ea ed as objec s:
Bac e ial esis ance genes compu ed ei he as an
independen uni o cons i u ing a combina ion
oge he wi h a pa icula subcellula eplica o ( o
ins ance a plamid). He e, we use he se o symbols
AR
i
o desc ibe genes encoding AR, whe e ideno es
he objec ’s iden i y. Should an AR gene is designa ed
in i s single o m (AR-like) i will be conside ed as a
genomic gene (i.e. p esen as a locus wi hin he
bac e ial hos genome) by he model, bu i i is
a ached o a pa icula subcellula eplica o hen i
will be conside ed o be pa o he subcellula
eplica o (i.e. ca ied by he subcellula eplica o ).
Subcellula eplica o s inside bac e ial cells such as
plasmids, in eg a i e-conjuga i e elemen s (ICE),
ansposons, o any o he gene ic elemen wi h sel -
eplica ion abili y. Fo he sake o simplici y, in his
i s e sion we only conside plasmid-like objec s,
which ou model compu es wi h he se o symbols
PL
i
. As p e iously indica ed abo e, he model
pe mi s simula ion o plasmids ca ying AR genes by
in oducing a egula exp ession ha de ine complex
objec s as ollows: le us o conside kdi e en
plasmids and jdi e en AR genes, hen a complex
objec o he combina ion o plasmids wi h genes
belongs o he ollowing exp ession (wi h λbeing an
emp y s ing in he absence o a name): (PL
i
+…+
PL
k
)(AR
i
+λ)…(AR
j
+λ). Fo example he s ings
“PL
1
”,“PL
1
-AR
1
”,“PL
1
-AR
2
”, and “PL
1
-AR
1
-AR
2
”
espec i ely co espond o objec s ep esen ing ou
di e en o ms o he same plasmid –“no ca ying
AR genes”,“ca ying gene AR
1
”,“ca ying gene AR
2
”
and “ca ying bo h AR genes”. Fo compu a ional
sake, he cu en e sion o ou P-sys em model
pe mi s only simula ion o plasmids ca ying up o
wo di e en AR genes.
Ex e nal chemicals and/o bioma e ials eleased in o
he en i onmen , including any kind o molecule
inhibi ing bac e ial g ow h such as an ibio ics and
biocides. We use he se o symbols A
i
o desc ibe
hese objec s.
Managemen Clocks a e objec s labeled wi h G
i
symbols used o pe iodically add objec s o speci ic
memb anes acco ding o ecosys em in luences
expec ed o be cyclic.
In he memb ane s uc u e μ he model implemen s
i e hie a chical EBs labeled as ECO,P
i
,RS
i
,H
i
and B
i
ECO is he skin EB ep esen ing he pe iphe al
ecosys em o ul ima e con aine o all popula ions
and en i onmen s.
P
i
is he second EB ype in he memb ane
amewo k le el ha designa es simula ion o
pa icula en i onmen al a eas o he ecosys em o
he sp ead o hos s and esis an bac e ia and
e en ually o he bac e ial pa hogens (inhabi able
spaces, a eas o ood acquisi ion, and o he e en ual
in as uc u es).
H
i
is a ype o EB on he hi d le el o nes ing wi hin
he memb ane amewo k hie a chy. We use H-like
EBs o de ine hos indi iduals ca ying mic obio a
(hos -speci ic assemblies o bac e ial
mic oo ganisms, de ined by i s mic obio a
composi ion).
RS
i
is ano he ype o EB, also on he hi d
amewo k le el, ha he model uses in o de o
abs ac (when sui able) he compu a ion o h ee
di e en ypes o physical o concep ual ese oi s.
Two o hese a e called Food and Wa e supplies
and a e used du ing he simula ion as ese oi -EBs
o hese esou ces. I is wo h o no e ha o
compu a ional sake ood and wa e a e no ea ed
as objec s bu as quan i a i e in e nal esou ces o
he P-sys em being necessa y o hos s´ li e ha
mus he e o e be pe iodically gene a ed by
managemen clocks and consumed by he simula ed
hos s acco ding o he a es s a ed by he use s in
he con igu a ion o ules. Once he hos popula ion
g ow h ou paces he a ailabili y o ood and wa e
he model ac i a es an in e nal mal husian ule ha
andomly kills (elimina es om he simula ion) a
numbe o hos s equi alen o he popula ion
su plus. The hi d RS-like EB is called “Sewage”and
i e e s o any body o wa e con eying all wa e -
ca ied was e (ei he na u al o an h opogenic) being
emo ed om a communi y. Sewage can also be
used o simula e he s ool emains (o ecal
d oppings) pe iodically eleased by H-like indi iduals
o he en i onmen . The h ee RS-like ypes o EBs
a e ep esen ed only once wi hin each (P-like)
en i onmen and al hough hey a e p o ide s o
wa e and ood hey may also con ain subs ances
and mic obial con amina ion eleased h ough he
d oppings o animal hos s and he con e sion o
dead animals in o ood. In o he wo ds, he use o
ese oi s allows he use o simula e supplies o
ood and wa e bu also ecycling o mic obio a
eleased in o he ecosys em by animal hos s (H-like
memb anes) du ing he inal ac o diges ion o o
Campos e al. Biology Di ec (2015) 10:41 Page 4 o 13
u n in o ood any o ganism ha dies o is p eda ed
by o he o ganisms.
B
ij
±
is he las EB le el o he hie a chy con empla ed
in ou model and i is used o simula e bac e ial
cells. Each cell has se e al a ibu es he e de ined as
ollows: he supe sc ip “plus/minus”is used o
indica e i he cell is g am-nega i e o g am posi i e;
he subsc ip iis used o ep esen cell popula ions
as lineages e m he e used o highligh ha he use
can design he simula ion o a mic obiome acco ding
o a common his o ical o sp ing o cells a any
axonomical le el (a lineage can he e o e e e o a
domain, phylum, class, o de , amily, genus, species
and clones, depending on he expe imen ); jis used o
assign wo o mo e cells o a pa icula o unique
communi y o cells o GEC. Fo ins ance, wi hin a
mic obiome, he use can de ine a subse o cells o
belong o GEC
1
and ano he subse o belong o
GEC
2
using he subsc ip di e en ia ion.
The dynamics, speci ici y and beha io o he dis inc
memb anes and objec s du ing he P-sys em simula ion
a e adminis a ed by a ini e se o ules (R) ixed o
each memb ane ha can be anked by o de o p io -
i ies (ρ). Pa icula ly, ou P-sys em model conside s
ules o he p ocesses o ansi ion, in e ac ions and so-
cial beha io , bi h, dea h, inac i a ion, and e olu ion o
he dis inc objec s and memb anes. The model also
conside s ules called speci ica ions o ecosys em e-
sou ce limi a ions in space and ime (as limi s in space,
o li e expec ancy). Following a e some examples o
ules desc ibed in he o mal de ini ion ollowed by he
memb ane compu ing communi y, and whe e o he
sake o simplici y, we omi p io i ies and s ochas ic
pa ame e s.
Example 1: ansi ion ules wi h mo emen o hie -
a chically adjacen egions:
“Subs ance Ai en e s bac e ial cells”o “Subs ance Ai
lea es bac e ial cells”.
½ui½
jk→½ui0
½
jkinsideðÞ
½ui
½
jk→½ui0½
jkou sideðÞ
Example 2: ansi ion ule wi h mo emen o win
adjacen egions:
“A bac e ial cell ansmi s a plasmid o o he membe s
o he same GEC”.
ui
½
j½
i→½
jui0
½
i
Example 3: ac i e memb ane ules wi h mo emen o
win adjacen egions:
“An indi idual hos ansmi s bac e ia o ano he
indi idual”.
½½ 
kj½
j→½
j½½
ki
He e, egion kis mo ed om egion j o egion i, ha
is a he same le el. Then, egions inside egion kcan be
mo ed acco ding o hei co esponding ules.
Example 4: ac i e memb ane ules o memb ane
di ision:
“G ow h o bac e ial cells j in egion i.”
½½ 
ji→½½
j½
ji
He e, he con en o egion jis copied oge he wi h
i s ules and all he memb anes i con ains in a hie a ch-
ical manne .
In summa y, ou P-sys em consis s o a memb ane
s uc u e composed o i e ypes o EBs and a wo king
alphabe Vo objec s whose in e ac i e eedback is de-
e mined by he se o ules assigned a e e y EB. This
amewo k can be g aphically ep esen ed as a Venn dia-
g am (Fig. 1). As shown in he igu e, he con aine dia-
g am (ECO) ep esen s he skin EB. Wi hin ECO, he
wo nex diag ams designa ed as –P
i
and P
j
– ep esen
wo P-like en i onmen al EBs ( he use can howe e de-
sign as many P-like EBs as equi ed). P-like EBs a e
allowed o con ain RS-like, H-like and B-like EBs ( ep e-
sen ed as diag ams o smalle size). RS-like EBs (desig-
na ed as i, j, k) ep esen ood, wa e and sewage
ese oi s and a e allowed o con ain B-like memb anes
(bu no H-like memb anes). H-like EBs can be dis in-
guished in sub ypes (social classes, species, e c.) using
subsc ip assigna ions. Fo example, in he igu e we
con empla e 3 popula ions (i, j, k) ha may be espec -
i ely composed o a numbe o indi iduals ( o ins ance
100, 50 and 150, e c.). Each H-like EB is allowed o con-
ain a numbe o in e nal B-like EBs (bu no RS-like
EBs) de ining i s in insic mic obio a. B-like EBs can be
placed no only wi hin RS-like and H-like EBs bu also
in P-like EBs and can be di e en ia ed in lineages o
which g am and GEC s a us can be assigned using sub-
and supe sc ip s. The s a us o G am posi i e o G am-
nega i e o ganisms is assigned using a supe sc ip wi h
wo s a es (minus and plus). In he igu e we obse e
ou sub ypes (i
_
,j,k,l) acco ding o he le subsc ip .
Those labeled wi h he le subsc ip jbelong he GEC-j
and hose labeled wi h he subsc ip kbelong o GEC-k.
Those ha ing he supe sc ip plus a e conside ed o be
g am-posi i e cells, and hose assigned he supe sc ip
minus a e g am-nega i e. Logically, popula ion size can
also be assigned o each lineage ( o example 10
9
cells
pe bac e ial lineage). The wo king alphabe is composed
o ou ypes o objec s (also di e en ia ed in sub ypes)
summa ized below he Venn diag am. In pa icula he
Campos e al. Biology Di ec (2015) 10:41 Page 5 o 13

igu e shows wo AR-like objec s (i, j) de ining wo di -
e en AR genes; ou A-like objec s (i, j, k, l) de ining
ou dis inc subs ances ( o ins ance wo an ibio ics and
wo insec icides wi h di e en p ope ies); eigh PL-like
objec s ep esen ing wo plasmids (i, j) each one wi h
ou possible s a es (wi hou AR genes, ca ying an AR
i
gene, an AR
j
gene o ca ying bo h AR genes); and ou
G-like objec s (i, j, k, l) ep esen ing managemen clocks.
AR-like and PL-like objec s a e es ic ed o B-like EBs
bu hey can mo e om a B-like EB in o ano he ( o
emula e ho izon al ans e e en s). No e, howe e , ha
AR objec s can be only ans e ed when ca ied by a
Fig. 1 P-sys em model o AR e olu ion in complex ecosys ems. Venn diag am ep esen a ion showing o he amewo k o memb anes and
ocabula y o objec s, on which ou P-sys em model is based; memb anes a e illus a ed as nes ed diag ams labeled a bo om acco ding o he
model´s code o symbols we use o e e ing memb anes; objec s a e also ep esen ed using symbols summa ized below he igu e; and ules
assigned o each memb ane a ea a e, o simplici y´s sake, indica ed as ex indica ions colo ed g een
Campos e al. Biology Di ec (2015) 10:41 Page 6 o 13
plasmid objec . A-like and G-like objec s a e allowed in all
EBs (excep ing ECO which is he ul ima e con aine ) as
hey ei he de ine subs ances expec ed o sp ead ac oss all
en i onmen s o pe iodical ac ions (in he case o G-like
managemen clocks) s a ed by he use . Finally, e e y EB is
assigned a se o speci ic ules (R) designed and uned wi h
he aim o go e n he dynamic o in e ac ions and e olu-
iona y e en s wi hin each EB acco ding o gi en p io i ies
(ρ), pa ame e s and condi ions indica ed by he use .
Resul s and discussion
In oducing ARES: simula o de ice co e and se e
implemen a ion
The P-sys em model p e iously de ined in Me hods was
implemen ed as a simula o so wa e, which was p o-
g ammed using he Ja a objec -o ien ed compu e p o-
g amming language [49] and P-lingua ounda ions [50].
In pa icula , he simula o eads an xml ile ha con-
ains he s a ing con igu a ion o a P-sys em and un
simula ion o he P-sys em case s udy con igu ed du ing
a use -de ined numbe o in e ac ions con empla ing
ou s eps pe i e a ion; 1) s ep o e olu ion whe e all
ules o e olu ion and in e ac ion a e applied; 2) s ep o
mo emen whe e all ules o mo emen apply; 3) s ep o
g ow h whe e all memb anes allowed o di ide do so.
A e each s ep, an upda ing p e-s ep p epa es he model
o he nex one. The ules applying o each egion ac
simul aneously, in a pa allel way, o all he objec s and
egions, and he s ochas ic beha io o he sys em is
achie ed by applying he ules acco ding o hei s o-
chas ic pa ame e s in a nai e p obabilis ic manne . Rules
apply i : a) he objec s in need o he ule applica ion a e
in he egion and b) he e is no ule wi h highe p io i y
ha uses common objec s. The ansi ion om one con-
igu a ion o he nex is ca ied ou by applying all he
ules a e e y egion in a non-de e minis ic maximal
pa allel mode; he sys em is always unning om one
con igu a ion o he nex and only hal s i no ule can be
applied. Hence, he hal ing con igu a ion con ains he
ou pu o he sys em and he inal con igu a ion. In
addi ion, s ochas ic pa ame e s o model he popula ion
dynamics o he sys em a e also in oduced in he ules.
The simula o has been ins alled wi hin an engine co e
wi hin a 4× 6 Co e Se e wi h Linux OS and 128 GB o
RAM, and has been coupled wi h he ollowing sub-
sys ems; 1) a MySQL Managemen Sys em o s o age o
P -sys em con igu a ions; 2) an engine o ou pu con e -
sion o CSV o ma ; 3) he ou pu ´s a chi e, which is an
eposi o y o ou pu olde s; 4) a se e sec ion o upload
aining u o ials; 5) a collec ion o sc ip s o s a is ical
analysis de eloped using he R p og amming language
[51]; 6) A on -end in e ace layou o manage all o he
sub-sys ems, p og ammed in a PHP amewo k on
La a el 4 ollowing he Model–View–Con olle pa e n
an a chi ec u al model whe e se e in e aces a e in e -
changeable [52]. This In as uc u e is wha we call ARES.
Managing ARES
Use s can simula e dynamics o AR e olu ion using
ARES o design a P-sys em model scena io adap ed o
he case s udy speci ied by he use and hen un simula-
ion o his scena io as many imes as necessa y co ec -
ing pa ame e s un il a ealis ic desc ip ion o he AR
p ocess is app oxima ed o alida ed i eal wo ld obse -
a ions a e a ailable. Use o ARES is ee, bu i s acces-
sion is passwo d p o ec ed in o de o allow he use s o
open and main ain a use accoun ha will needed o
s o e and un model p ojec s in p i a e session (a simu-
la ion may las hou s o e en days depending on he
complexi y o a P-sys em scena io). ARES is managed
ia an easy- o-use in e ace ha implemen s a cen al-
ized menu (Fig. 2a) o accessing he sys em o o ms he
use need o sequen ially comple e in o de o in oduce
he s a ing con igu a ion o a P-sys em scena io, un a
simula ion, and access he esul s. All menu- o ms access-
ible wi h his menu can be na iga ed back and o h o
edi ing he P-sys em con igu a ion, change o add EBs,
objec s and ules whe e o when necessa y. A scheme o
he whole ARES in as uc u e and he wo k low o
con igu a ion and simula ion o P-sys em scena ios is
depic ed on Fig. 2b. The usual p ocedu e can be syn he-
ized in he ollowing s eps. The o m designa ed as
“ECO”mus be i s accessed ( ia menu) and comple ed o
c ea e he P basal skin EB; hen “ENVIRONMENTS”has
o be accessed o con igu e as many P-like (en i onmen-
al) EBs as needed wi hin ECO; nex , “RESERVOIRS”and
“HOSTS”mus be illed o con igu e RS-like ( ese oi s)
and H-like (hos s) EBs wi hin he p e iously c ea ed P
EBs; a e his, “MICROBIOMES”mus be used o con ig-
u e a se ies o B-like (bac e ial) EBs ha can be ei he
placed wi hin he p e iously c ea ed P-, RS- and H-like
EBs; hen “OBJECTS”mus be used o c ea e as many as
PL- (plasmids), AR- (AR genes), A-like (an ibio ics and/
o o he subs ances), and G-like (clocks) objec s as e-
qui ed wi hin he p e iously c ea ed EBs (excep o
ECO, since i is he skin memb ane); inally, he o ms
“SPECIFICATIONS”and “RULES”mus be comple ed o
s a e he ules assigned o each EB by selec ing hem
om a lis o p e-designed ules p o ided in an unde -
s andable and gene alized way allowing he use o
choose and une ules wi h he alues and he pa ame e s
needed o app oxima e he equencies, beha io s, condi-
ions and p io i ies ha go e n he dynamic o in e ac-
ions among he di e en memb anes and objec s o he
P-sys em model o be simula ed.
Once he P-sys em s a ing con igu a ion has been de-
ined, i is au oma ically w i en o an xml ile (which is
he inpu in o he simula o ), which is s o ed in ARES.
Campos e al. Biology Di ec (2015) 10:41 Page 7 o 13
Fig. 2 ARES in e ace and se e o ganiza ion. aSc eensho o he ARES in e ace. The in e ace implemen s a menu ha gi es access o he
dis inc se e o ms ha apply o con igu a ion, s o age and simula ion o P-sys em model scena ios. A he bo om o he in e ace he use
can access o he suppo sec ions o managing ARES o o s a is ical in e oga ion o he ou pu gene a ed by he simula o de ice. bARES se e
scheme and wo k low o c ea ion, edi ion and simula ion o P-sys em model scena ios
Campos e al. Biology Di ec (2015) 10:41 Page 8 o 13
The o m “RUN”is an in e ace ha p o ides accession
o he simula o engine co e accompanied by a lis o all
xml iles eady o simula ion. The use only needs o se-
lec an xml ile om such a lis , hen de e mine he
numbe o i e a ions he simula ion will las (each i e -
a ion is se o be co espond o one day) and un he
simula ion. I is wo h o no e ha xml iles can be
expo ed o impo ed om he use ’s PC as con enience,
using he “RUN”in e ace.
Once he simula ion s a s, ARES au oma ically c ea es
a olde (labeled wi h he name gi en o he P-sys em
du ing he con igu a ion) in he ou pu s’a chi e assigned
o his pa icula simula ion and hen gene a es he xml
inpu ile ha is placed wi hin his oupu olde . When he
simula ion inishes he sys em i s deli e s a aw ou pu ,
which is a plain ile con aining he sampling coun s pe i -
e a ion o all simula ed objec s and EBs. The aw ou pu is
di icul o manage because o he de aul o ma i en-
closes. To o e come his di icul y, ARES implemen s an
ou pu con e e engine ha p ocesses and spli s he aw
ou pu in o a se o 5 o ma ed cs iles appoin ed
as ECO-like.cs , P-like.cs , RS-like.cs , H-like.cs , B-like.cs .
These iles cons i u e oge he he ou pu ha ARES de-
li e s a e simula ion o he use in o he ou pu ’s a ch-
i e ( he ou pu ’s a chi e can be accessed using he
“OUTPUTS” ab in he ARES menu). Each cs con ains
he coun s o each EB and objec sampled a he
P-sys em EB le el e e ed in he ile name. Fo ins ance,
ECO-like.cs has he coun s o all EB and objec s sampled
a he ecosys em le el, H-like.cs has he coun s o all
objec s and EBs sampled in he H-like EBs bu no in
hose o highe le els (i.e. ECO and P-like); and B-like.cs
ile has he coun s o all objec s sampled in B-like EBs. In
each cs , i e a ions co espond wi h he ows and a e
o ganized in ascending o de (being he s a ing
con igu a ions he i s coun in he ile) and memb anes
and objec s co espond wi h he columns.
ARES also o e s o he sec ions o use suppo ,
which can be accessed ia he submenu a ailable a he
oo e o he ARES in e ace. 3 o hese sec ions called
“R-TOOLS”,“TUTORIALS”,and “AVAILABLE RULES”
a e o pa icula in e es .The i s sec ion (R-TOOLS)
gi es access o an in e ace (also managed ia menu)
o e ing he use s di e en sc ip s mainly bu no exclu-
si ely de eloped in “R”, o downs eam in e oga ion o
cs ou pu s. No e howe e ha he use o R-TOOLS is
no a manda o y ask as cs s a e open plain iles ha
can be p ocessed using any o he ool o s a is ical pack-
ages such as Excel, Gnuplo , Ma lab and Ma hema ica,
e c. The second sec ion (“TUTORIALS”) is a eposi o y
whe e use s can upload and download u o ials o P-
sys em con igu a ion and managemen o ARES (see also
he sec ion below, “Tu o ials and aining ma e ial”).
Finally he hi d sec ion (“AVAILABLE RULES”)isa
sec ion whe e we summa ize all p e-designed ules o da e
a ailable o each ype o EB in o de o le he use o
make p elimina y e alua ions o he ules o ake o a pa -
icula simula ion be o e c ea ing he s a ing con igu a ion
o he P-sys em. This sec ion also includes a o m o use s
o make i necessa y any speci ic sugges ion o he
implemen a ion o new ules no ye a ailable.
Tu o ials and aining ma e ial
Al hough managemen o ARES is qui e in ui i e, he
design and p epa a ion o he s a ing con igu a ion o a
P-sys em model scena io can be an a duous ask o e-
sea che s no amilia ized wi h memb ane compu ing
(con igu a ion o i ems, assigna ion o ules, e c.). Taking
his in o p ima y conside a ion, we ha e p epa ed wo
u o ials aimed o gi e he eade some aining ma e ial
ha can be downloaded om he “TUTORIALS”sec ion
o ARES unde he labels “Nosocomial Scena io”and
“Two cock oach a ms” espec i ely. The i s u o ial
con empla es a simpli ied nosocomial scena io p o ided
wi h he sole objec i e o allowing he use o ake he
i s s eps in lea ning how o con igu e an exempla y P-
sys em c ea ing wo memb ane en i onmen s (a commu-
ni y and a hospi al en i onmen ), hos EBs ( o ins ance,
pa ien s), mic obial communi ies wi hin hos s, composed
o dis inc bac e ial EBs and plasmids, ca ying AR genes
objec s wi hin he bac e ial EBs. The u o ial also exempli-
ies how o c ea e clocks o in oduce o he objec s such
as an ibio ics in he simula ion o how o con igu e and
une a basic package o ules. The second u o ial con em-
pla es a mo e elabo a ed scena io ha can be add essed
a e comple ing he i s u o ial. This u o ial speci ically
ocuses on he simula ion o wo popula ions o B. ge ma-
nica (a model insec o ganism able o implan an in es-
inal mic obio a simila o ha o humans [53–56])
espec i ely emplaced in wo sepa a e cages wi h he pos-
sibili y o mig a ing om one o he o he . These wo
boxes concep ually ep esen en i onmen al EBs o hos-
pi al and u ban-communi y indi iduals, he e designa ed
as P
1
and P
2
. Cock oaches o bo h a ms a e hos s (H-like
EBs) ca ying he same in es inal mic obio a, which ac-
co ding o Ca asco e al. [56] is p edominan ly composed
o eigh bac e ial lineages (B-like EBs). Fou o hese cellu-
la lineages will be simula ed as G am-nega i e, while he
ou o he will be G am-posi i e. All bac e ial cells o all
lineages a e allowed o ca y h ee dis inc ypes o in a-
cellula plasmids (PL
1
,PL
2
and PL
3
) capable o ho izon al
ans e . One o hese (PL
1
) is ca ie o an AR gene
(AR
1
) con e ing esis ance o a g am-nega i e speci ic
an ibio ic designa ed as “A
1
”while ano he plasmid ype
(PL
2
) ca ies an AR gene (AR
2
) o e ing esis ance o a
g am-posi i e speci ic an ibio ic labeled as A
2
. In he
s a ing con igu a ion, he hi d plasmid ype (PL
3
) does
no ca y AR genes bu du ing he cou se o he
Campos e al. Biology Di ec (2015) 10:41 Page 9 o 13