►
From YouTube: DevoWorm #44: Workshop(s) recap/updates, Sodaplay tensegrity, Morphogenetic Tug of Wars and Locks
Description
Recap of the Morphing matter, Learning on Graphs, and MeMoDevo workshops, Mayukh Deb's Gradient article, Sodaconstructor (open-source Sodaplay platform) as a basis for tensegrity networks. Synthetic Biology as a testing ground for Development and Morphogenesis. Papers on a viscoplastic lock for progressive body-axis elongation and tugs-of-war between filament treadmilling and myosin-induced contractility. Attendees: Richard Gordon, Susan Crawford-Young, and Bradly Alicea
A
A
A
A
A
B
A
That
one
you
start
reading
it,
it
makes
sense
until
you
get
into
the
middle
of
it
and
it's
like
you
left
half
of
it
out,
I,
don't
know,
there's
elasticity
and
just
kind
of
shear
stress
and
you
I,
don't
know
whether
that's
the
elasticity
and
shear
stress
of
the
whole
thing
and
you
have
to
measure
it.
Yeah
I
think
that's
what
you
do,
but
I'm
actually.
B
B
B
B
A
B
A
B
A
A
A
A
B
B
A
A
B
A
B
B
B
A
A
Okay,
physics
of
morphing
matter,
December
12th,
the
14th
topics
were
active
algorithms
for
morphable
matter,
reconfiguring
meta
structures
from
Wave,
Control
to
metado
intelligence,
safe
morphing,
morphing
and
chirality,
switching
directed
by
spatially,
encoded,
Liquid,
Crystal,
elastomer,
microparticles,
inflatable
morphing
matter
using
bio,
hybrid
robotics
to
understand,
structure
and
function,
relationships
of
the
heart,
computational,
design
of
soft
functional,
composite
structures
and
robots
and
liquids
and
elasticity,
and
then
I
missed
mechanics
guided
through
the
assembly
and
I
love.
This
shape,
multi-step
shape
morphing,
so
I'm
upset
about
some
of
this.
B
A
B
A
B
Yeah,
they
have
a
whole
agenda
here.
Yeah
this
shape.
Morphinger
was
really
interesting
because
I
went
to
another
Conference
during
that
same
time
where
they
talked
about
some
of
the
work,
people
are
doing
with
inflating
different,
like
building
balloons
or
some
sort
of
inflatable
structures
that
you
can
actually
mimic
morphogenesis
with.
So
you
can
actually
mimic
like
the
plant
body
or
things
that
are
taking
shape
and
development,
and
you
can
look
at
the
structural
stability
and
and
how
they
kind
of
unfold,
how
they
move,
how
they
buckle.
And
then,
oh.
B
A
B
B
A
B
A
B
A
A
So
these
are
inflatable
structures
up
here
and
they've
actually
calculated
how
they
inflate
they've,
made
them
out
of
triangles
and
carefully
went
over.
This
they've
made
them
out
of
triangles
and
they've
built
them
and
they've
found
out
how
that
how
they
work
and
they
plotted
where
they
will
properly
inflate
and
where
they
won't.
B
A
Three-Dimensional
structures:
okay,
yeah
three-dimensional
structures,
sort
of
organic
experience,
so
they
did
that.
Oh
maybe
this
smell
as
the
heart.
This
was
good
about
the
heart.
A
heart
is
a
a
sort
of
a
sigmoid
shape.
It's
a
it's,
not
a,
not
a
hollow,
it's
more
of
a
wrapped
wrapped
piece,
so
they
potent
toilet
paper.
It's
a
wrapped
muscle.
B
A
B
A
B
B
A
Skeleton
of
a
heart
and
put
rat
myocardiac
Senate
and
get
pumped
and
they
found
out
you
had
to
make
the
spirals
not
just
straight
across
but
at
an
angle,
or
else
it
didn't
pump
the
way
it
should,
and
then
they
made
a
fish
that
swam
here's
the
fish
and
I
know
the
guy
said
he
visited
aquariums.
So
maybe
that's
why?
Why
all
of
that
anyway
and
they're
working
on
on
making
the
whole
heart
actually.
A
B
A
It
actually
works
as
a
gripper.
Okay,
if
you
make
it
like
this,
I
was
more
interested
in
now,
there's
tubes
in
our
body
like
some
of
these
things,
so
they
go
from
this
to
to
that
and
they
they
will
bend
as
well,
so
they're
making
objects
that
move
and
they
got
us
this
line,
inchworm
thing
to
to
move
as
well.
B
A
A
B
B
A
B
A
A
This
is
this:
is
the
experiment
using
these
these
things,
and
it's
really
neat.
You
can
put
these
on
the
end
of
another
one
and
these
these
wrinkle
and
I
take
hold
of
whatever
that
is
and
then
lift
she
made
a
grab
and
hold
and
pull
on
these
singles
fingers
here.
Curl,
okay
and
a
thickness
thing
curl
at
different
times,
due
to
the
thickness
of
the
material.
B
A
B
B
B
B
A
B
A
B
A
A
A
B
Yeah,
it's
good,
so,
okay,
I
have
a
bunch
of
things
actually
to
to
go
through.
I
went
to
two
meetings
last
week
and
they
were
during
this
time
that
you're
the
meeting
you
just
showed
us,
but
the
second.
The
meeting
that
I
talked
about
Nemo
DeVos
actually
had
some
very
similar
topics
to
this.
So
I'm
going
to
go
through
a
little
bit
of
that
and
show
you,
okay.
B
All
right
so
first
thing
I
want
to
talk
about.
Was
this
new
video
on
the
diva
one
YouTube
channel?
This
is
on
embryo
networks
and
computational
development.
So
this
is
about
an
hour-long
talk.
That
summarizes
a
lot
of
the
work.
That's
been
done
on
the
group
on
embryo
networks
and
some
of
the
network
structures
that
we've
written
papers
on,
and
things
like
that.
So
it
kind
of
goes
over
the
last
five
years
of
that
work.
I.
B
Guess
it's
been
about
five
years,
so
this
goes
into
this
idea
of
embryo
networks,
which
are
where
you
have
cells
or
things
within
cells
that
form
networks,
and
you
can
build
embryo
Network.
So
you
can
build
Network
models
of
the
embryo
at
different
points
in
development.
You
can
build
connectomes,
which
are
where
you
model
neurons
in
the
nervous
system,
and
you
can
build
other
types
of
networks
that
are,
you
know,
have
other
functions.
B
B
B
B
I
think
that's
the
wrong
picture,
yeah,
okay,
so
those
are
all
available.
The
next
thing
I
want
to
talk
about
is
this.
This
was
my
Europe
Deb,
of
course
worked
with
us
two
years
ago
on
Google
summer
of
code,
and
he
has
been
a
maintainer
with
medieval
learn
platform,
and
so
he
recently
wrote
an
article
in
the
gradient,
which
is
a
machine
learning.
It's
like
out
of
a
popular
Journal.
It's
you
know.
They
publish
articles
every,
maybe
two
weeks
on
different
topics
of
machine
learning,
and
so
he
he's
been.
B
He
is
finally
able
to
get
his
article
out,
so
mayork
has
been
active,
doing
things
with
the
startup
LF
Alice,
so
he's
been
working
on
neural
networks
there
and
he's
been
doing
other
things
he's
been
he's
been
a
success
story
of
our
Google
summer
code
program.
You
agree
yeah
and
he
wrote
this
article
learning
to
make
the
right
mistakes
a
brief
comparison
between
human
perception
and
multimodal
learning
machines,
and
so
this
is
a
nice
article
kind
of
goes
over.
B
You
know
multi-modal
language
models,
which
are
these,
this
class
of
models
where
they,
if
you've
heard
of
large
language
models,
that's
kind
of
what
they
are,
but
multi-modal
language
models
are
predict
like.
If
you
give
it
a
prompt,
it
gives
you
a
text
string
that
can
be
predicted.
There
are
these
next
token
predictors,
which
is
the
mechanism
behind
along
these
models.
So
gpt3
is
a
large
language
model
and
multimodal
language
models
are
an
attempt
to
make
such
language
models
perceive
the
world
in
a
way.
B
That's
one
step
closer
to
humans,
so
you
can
you
know
if
you
have
something
like
a
resnet
or
GPT
Neo,
those
are
specialized
for
vision
or
language,
but
not
both,
and
so,
if
you
use
a
resonant
and
you
give
it
visual
information
like
a
database
of
images,
you
can
use
it
to
predict
the
resonant
to
predict
the
images
and
then
the
larger
language
model
to
put
up
the
text
or
the
word
streams,
and
you
can
have
a
multimodal
model
where
you
have
images
and
words
I.
Think
it's
emerging
there
together.
B
So
this
is
an
example
here,
where
you
have
a
picture
of
a
bird
and
you
encode
the
image,
and
then
you
have
the
string,
which
is
this
is
awesome.
It
could
be
potentially
like
a
description
of
the
bird
or
some
other
semantic
information
use
a
text
tokenizer
and
then
this
forms
together.
This
forms
an
embedding
and
then
this
goes
to
a
pre-trained
language
model
and
then
there's
an
output
and
it
classifies
it
as
a
bird.
B
So
he's
got
a
lot
of
and
then
he
has
some
closing
words.
So
yeah
check
that
out
it's
a
nice
article,
he
wrote
so
the
next
thing
I'm
going
to
talk
about
are
these
models
that
comes
up
actually
I'll
start
here.
These
are
called
soda
Constructor
models
and
these
are
pretty
old,
actually
I,
remember
playing
with
them
about
15
years
ago.
They're
these
physics
models
that
are
like
they
have
these
points
and
these
struts
and
the
idea
is
that
these
points
are
weight,
masses
or
or
just
Mass,
Point
masses.
B
So
you
know
they're
they
at
one
time
there
was
this
tool
called
soda.
Constructor
and
soda
Constructor
was
a
very
popular
platform
to
build
your
own
models
on
and
they
had
tournaments
where
people
would
build
models
that
would
do
all
sorts
of
things
they
had
Walkers
like
this,
and
they
had
different
types
of
models
that
you
know
would
twirl
around
and
locomote
and
all
sorts
of
interesting
ways
and
move
around.
In
these
physics.
B
A
standard
physics
simulation,
it's
a
very
simple
physics
simulation,
but
it
has
these
basic
elements
and
what
you
can
do
with
these
models
is
you
can
setting
a
number
of
parameters
on
these,
so
the
simulation
is
basically
having
these
points
in
space
having
these
Springs
or
spring
masses
that
join
between
them,
these
struts
and
then
you're
able
to
simulate
them
with
a
force,
and
so
this
is
a
the
spring
you
can
play
around
with
this.
The
sidebar,
which
shows
these
little
spring
masses,
are
a
certain
frequency.
B
If
you
play
around
with
them
a
little
bit,
you
can
get
it
to
do
different
things,
so
it's
slowed
down
a
little
bit
all
right
now,
I
touched
it
and
it
start
moving.
It
starts
to
move
it
moved
up
in
the
air,
and
now
it's
kind
of
walking
weird
a
little
bit
more.
The
gate
is
a
little
bit
different,
so
you
can
play
around
with
the
parameters
like
that.
So
it's
a
really
interesting
tool.
B
Now
soda
Constructor
has
been
off
the
internet
for
a
while,
but
people
have
built
the
code
is
open
source
and
you
can
build
your
own
Constructor
platform.
So
this
is
open
Constructor.
This
is
by
Peter
fidelman,
and
you
know
his
kind
of
revived
the
soda
Constructor
code
and
put
up
a
little
interface
for
that.
B
B
So
this
screenshot
I
have
here
is
this
is
actually
of
a
different
thing.
It's
a
genetic
car
which
is
built
on
the
same
principle.
That's
not
really
useful
in
this
context,
but
it
shows
you
that
yeah.
So
this
is
my
search
on
soda
Constructor
here
where
you
have.
This
is
the
actual
it
was
released
in
2000
and
it
was
a
Java
based
physics
engine.
B
B
There's
this
article,
it's
an
interview
with
Ed
Burton,
it's
evolving
out
of
the
virtual
mud.
This
talks
about
the
different
types
of
soda
Constructors
that
people
had
built
over
the
years.
There's
all
this.
You
know
these
blockers
and
different
types
of
spinning
structures
and
things
like
that.
Here's
also
this.
What
is
this
one?
Well,
this
is
soda
Constructor
physics.
B
This
is
actually
an
activity,
for
this
is
actually
an
educational
tutorial
for
I
guess:
high
school
students.
It's
pretty
easy
because
you
can
build
these
soda
Constructors
pretty
easily,
and
it
just
demonstrates
some
of
the
creatures.
So
you
know
you
can
build
them
like
kind
of
like
Vehicles
kind
of
like
it
was
like
10
seconds.
A
B
He
runs
them
on
the
beach
and
then
other
ones.
This
is
in
the
city,
but
basically
they
they're
they're
the
same
type
of
structure
as
the
soda
play
or
the
soda
Constructor
creatures,
but
they
actually
walk
in
their
physical
models,
they're
not
in
a
computer.
So
you've
used
these
same
principles
to
build
these
things,
and
some
of
these
creatures
even
have
a
stomach
without
conventional
animal
muscles,
so
he's
able
to
articulate
structures
in
the
animal
or
in
the
made
up
animal
without
muscles.
B
So
this
is
interesting
and
then
this
is
I.
Guess
the
winging
bottle
propelled
stomach
of
this
organism.
You
know
anime,
it's
non-organism.
It's
a
model,
they
gave
it
a
a
Latin
species,
name
in
in
genus
name,
and
then
this
is
so.
This
is
the
structure
here
and
it's
a
lot
like
some
of
the
structures
that
we've
talked
about
with
respect
to
some
of
these
soft
active
matter
models,
and
then
this
is
a
kinetic
sculpture
here.
This
is
a
simulated
strand
beast
and
then
this
is
an
example
of
the
model.
B
B
B
Oh,
this
is
a
lecture
Darwin
Darwinism
on
a
desktop
so
to
play
in
the
evolution
of
a
digital
world.
So
this
is
something
Ed
Burton
delivered
in
2005
I
was
a
lecture
on
some
of
the
stuff,
so
there's
a
strong
tie-in
with
darwinian
Evolution
and
some
of
these
things
that
are
being
built.
You
know
there's
a
lot
of
variation
in
these
designs
and
you
know,
being
able
to
construct
them
based
on.
Just
kind
of
function
is
really
interesting
stuff
that
emerges
from
this.
B
So
that
brings
me
to
what
the
actual
things
that
I
attended
last
week,
which
were
actually
two
things.
So
the
first
thing
was
this
learning
on
graphs
conference,
and
this
was
a
interesting
conference
on
graph
neural
networks
and
I
know.
Last
summer
we
did
some
things
on
graph,
neural
networks
and
gsoc,
and
we've
now
Incorporated
that
into
the
devil,
medieval
and
repository.
B
So
if
you
want
to
check
out
what
we
did
last
summer
with
respect
to
this,
go
to
the
GitHub
Organization
for
Diva
learn
and
look
at
the
devograph
repository
a
diva
graph.
Repository
has
a
lot
of
that
software
in
it
hasn't
been
incorporated
into
a
release,
a
diva
learn,
but
that's
that's
coming,
maybe
then,
in
the
new
year.
So
there's
this
learning
on
graphs
conference.
This
is
a
really
good
conference.
It
went
over
a
lot
of
things
with
respect
to
graph
neural
networks,
machine
learning
on
graphs
and
geometry,
and
so
it's
really.
B
It
was
really
interesting
talks.
It
was
all
live
streamed
on
YouTube
the
the
lectures
are
available
on
this
live
stream.
So
if
you
go
to
their
YouTube
channel,
it's
there,
they
also
had
some
tutorials,
which
I
did
not
attend,
but
I
think
they'll
be
also
putting
those
up
on
YouTube
eventually,
so
that
there's
some
tutorials
on
how
to
build
graph
neural
networks
using
like
a
collab
notebook.
So
if
you
want
to
follow
along
you,
can
you
know
if
you
know
how
to
use
collab
notebooks
you
can
follow
along
with
the
tutorials.
B
You
can
build
your
own
graph,
neural
networks.
There
were
some
really
interesting
tops.
A
lot
of
it
was
overlap
with
the
neurops
conference,
which
is
the
conference
main
conference
for
machine
learning
and
deep
learning.
There
were
a
lot
of
interesting
talks
on
like
sort
of
the
applications
of
graph
neural
networks
to
medicine
and
to
biology.
B
If
we
take
a
look
at
I,
don't
know
if
they
have
some
of
these
up
I,
don't
think
they
have
a
good
list
of
talks,
but
it's
yeah.
There
were
a
lot
of
interesting
talks
on
so
some
of
them
were
on
like
medicine.
Some
of
them
were
on
biology.
Some
of
them,
though,
were
on
like
the
methods
linking
graph
neural
networks
to
like
category
Theory
or
thinking
about
some
of
the
theoretical
issues
in
graph
neural
networks.
B
Thinking
about
some
of
the
successful
techniques,
so
one
of
the
successful
techniques
for
constructing
graph
neural
networks
are
message,
passing
algorithms
and
but
sometimes
for
some
applications.
That's
not
sufficient,
so
you
know
they.
They
considered
like
different
application
domains
and
really
you
know
there
is
a
I
think,
a
strong
push
to
to
apply
graph
neural
networks
to
biology
and
some
of
these
other
areas.
A
B
Yeah,
well,
they
they
just
talked
about
graphs.
So
the
idea
is
that
you
create
these.
You
use
like
a
data
set
an
input
data
set.
It
could
be
like
a
microscopy
data
set
and
you
build
these
graph
embeddings
from
the
data.
So
if
you
have
a
bunch
of
points
that
represent
your
features,
then
you
have
to
infer
graphs
from
that
and
the
graphs
then
are
evaluated
you
can
you
can
do
whatever
you
want
with
the
embeddings
they're?
Also
not
embedding
approaches.
B
B
There
are
other
ways
of
doing
it
and
I
don't
know
which
is
most
appropriate
for
that
problem
or
for
some
of
the
problems
we
have
but
I
mean
I
know
like
use
the
embedding
approach
and
it
worked
okay,
but
there
may
be
better
approaches
and
so
I'm
going
to
be
preparing
a
blog
post
on
this
in
the
near
future
to
kind
of
go
over.
Some
of
these
things
that
they
talked
about
so
we'll
yeah
stay
tuned.
For
that.
B
A
B
This
is
the
YouTube
channel
learning
on
graphs
just
so
we
have
that.
Then
they
attended
the
memo,
Devo
Symposium,
which
is
mechanics,
orphogenesis
development
and
evolution,
and
this
has
actually
hosted
in
Paris.
So
this
is
the
network
of
people
who
attended
the
workshop.
This
is
the
neuroanatomy
GitHub,
so
the
neuroanatomy
GitHub
they
have
a
little
application
that
plots
out
the
attendees
based
on
their
interests
in
into
these
Networks.
B
So
devorin
group
is
here
in
this
group
and
then
there's
another
group
of
people
interested
in
I.
Don't
know
what
this
sub
network
is.
It's
so
I
mean
this
is
just
kind
of
a
cute
thing:
it's
not
really
for
analysis,
but
some
of
the
things
they
did
at
this
Workshop
really
interesting
talk.
So
this
was
in
took
place
in
Paris.
They
had
like
a
morning
session,
which
was
in
person
and
then
the
afternoon
session,
which
was
hosted
in
on
well.
B
B
As
a
selective
pressure
for
embryo
evolution,
they
had
also
some
work
on
Plants,
so
it
wasn't
just
animals,
they
had
a
number
of
different
species,
there's
a
very
in
terms
of
speech
species.
It
was
very
diverse
and
went
from
small
microbes
to
like
different
Collective
collectives
of
microbes
to
yeasts,
to
animals
of
different
orders
to
you
know,
plants
different
types
of
plants,
so
there
was
all
sorts
of
stuff
going
on
there.
There
are
a
lot
of
things
on
mechanics
on
blastocyst
morphogenesis,
there's
some
stuff
on
info.
B
B
B
There
are
some
things
on
fly
development,
which
was
actually
a
major
theme
doing
things
at
drosophila,
which
was
it's
a
very
nice
model
system,
because
there's
a
lot
of
changes
in
shape.
There's
a
lot
of
folding
buckling
there
different
types
of
features
in
the
drosophila
embryo
that
are
really
good
for
studying
this.
There
are
also
other
things
looking
at
different
Canal
networks
and
Jellyfish.
Looking
at
looping,
Network
morphogenesis
things
like
that,
so
they're
in
jellyfish.
B
They
also
provide
insights
into
some
Network
morphogenesis,
which
is
the
formation
of
like
nerve
networks
and
other
types
of
phenotypic
networks,
and
then
the
second
day
there
was
a
lot
of
interesting
stuff
on,
like
the
the
development
of
the
face
or
the
or
the
human
face,
they're
actually
using
a
human
face.
They
were
looking
at
the
morphology
of
the
skull
and
how
it
changes
and
they
were
using
something
called
anatomical
networks.
So
this
is.
A
B
So
there
are
some
very
interesting
things
in
primates
that
are
going
on
with
face
shape
Evolution,
then
there
are
some
things
about
biological
shape,
eggs
bodies
and
beaks,
and
this
talk
actually
highlighted
some
of
the
applications
of
of
dynamical
systems.
Theory,
especially
lagrangian,
coherent
structures,
just
some
of
the
developmental
trajectories
that
are
occurring
in
beak
development
shape
and
propulsion
among
green
algae.
B
So
this
was
where
they
were
looking
at
LG,
they
were
looking
at
very
a
very
small
number
of
cells
in
these
algal
colonies
and
looking
at
some
of
the
things
that
are
going
on
with
the
shape
of
them
and
how
they
use
their,
you
know
small
Colony
size
how
they
use
that
shape
to
move
around
their
environment.
Then
there
were
this
modeling
organogenesis
from
biological
first
principles,
so
this
is
more
of
a
theoretical
talk.
Plant
root
plant
growth
against
mechanical
obstacles.
B
So
this
is
kind
of
an
interesting
model
of
how
physics,
the
physics
of
networks
and
root
networks
overcomes
different
resistances
in
the
environment.
Then
this
one
was,
of
course,
on
this
work
on
yeast.
So
there's
this
work
that
that's
ongoing,
that
look
at
yeast
as
sort
of
a
model
for
the
precursors
of
multicellularity,
so
yeast
will
Clump
into
these
groups
into
these
little
clusters
and
they
do
this
sort
of
in
a
loose
Association.
B
B
B
B
B
B
I
believe
it
was
in
yeah
this
last
past
month
in
November.
So
this
article
is
scaling
of
complexity
and
synthetic
developmental
biology,
and
what
they're
trying
to
do
is
they're,
trying
to
use
synthetic
biology
tubes
or
inform
developmental
biology
and
we'll
see
what
that
looks
like
in
a
little
bit.
B
So
the
application
is
synthetic
biology.
Approaches
to
study
development
opens
a
possibility
to
build
and
manipulate
developmental
processes
to
understand
them.
Better
researchers
have
reconstituted
fundamental
developmental
processes
such
as
cell
patterning
and
sorting
by
engineering
circuits
in
vitro.
Moreover,
new
tools
have
been
created
that
allow
for
control
of
Developmental
processes
and
more
complex
organoids
and
embryos.
B
So
some
of
the
stuff
we
talked
about
earlier
in
the
meeting
on
you
controlling
shape,
using
artificial
structure,
soft
active
materials,
but
also
some
of
the
things
we've
talked
about
in
past
meetings
on
organoids
and
even
some
embryos.
So,
if
you're
doing
this
with
genetic
control
and
embryos,
where
you're
doing
like
genetic
engineering
or
even
like
Force,
you
know
manipulating
the
forces
of
the
environment
of
the
embryo,
these
are
all
synthetic
approaches.
B
Synthetic
approaches
allow
testing,
of
which
components
are
sufficient
to
reproduce
the
developmental
process
and
under
what
conditions,
as
well
as
what
effect
perturbations
have
in
other
processes.
So
you
know,
when
you
have
these
model
systems,
you
can
look
at
what's
sufficient
to
reproduce
a
developmental
process.
A
B
Know
maybe
we
don't
really
know
if
something
is
necessary
or
something
is
sufficient.
We
have
to
test
this
through
an
experimental
means,
and
so,
at
least
in
this
case
we
can
see
what
components
are
sufficient.
It
doesn't
mean
that
they're
necessary,
but
that
if
they're
present,
then
that's
something
that
is
going
to
have
an
effect
on
the
system
and
so
then
also
when
we
have
a
perturbation
in
development.
B
We
envision
that
the
future
of
synthetic
developmental
biology
requires
an
increase
in
the
diversity
of
available
tools
and
further
efforts
to
combine
multiple
developmental
processes
into
one
system.
So
there's
no
one
model
system
that
actually
does
all
these
things.
We
need
to
go
to
different
model
systems
and
use
those
as
inspiration.
So
one
outstanding
approach
has
been
the
introduction
of
synthetic
Gene
circuits
to
reconstitute
cell
communication
and
tissue
patterning.
B
So,
for
example,
there's
a
minimal
symmetry
breaking
system.
That's
been
engineered
to
reconstitute
lateral,
inhibition
based
on
cell
cell
contact,
their
membrane
Brown
like
ligand,
receptor
pair
Notch,
Delta
and
Notch,
and
so
this
is
a
classic
system
in
developmental
biology
where
they
introduce
these
membrane
bone.
Ligand,
receptor
Pairs,
and
this
constitutes
a
system
where
it
controls
the
fate
of
cells.
B
B
System
and
it's
not
something
you
can
do
for
an
entire
embryo,
so
this
is
something
that
we
you
know.
If
we
want
to
look
at
something,
that's
embryo
wide
or
maybe
a
response
to
physical
forces,
we
may
need
to
use
another
model
system,
so
long-range
communication
mechanisms
have
been
explored,
doing
achieving
vitro
patterning.
One
example
is
morphogenic
systems
where
cells
respond
in
a
concentration,
dependent
manner
to
diffusable
signaling
molecules.
B
We've
people
will
use
reaction,
diffusion
systems
which
we've
talked
about
in
the
meetings.
It
gives
us
a
better
understanding
of
conditions
under
which
a
short
range
activation
and
a
long-range
inhibitor
can
generate
complex
patterns,
and
these
examples,
full
Gene
circuits,
were
engineered
into
cells.
So
you
can't
just
use
the
gene
circuit
as
a
sort
of
a
simple
switch.
You
have
to
engineer
them
into
cells
and
use
them
in
as
part
of
the
cellular
machinery
showing
that
the
components
were
used
as
sufficient
used
were
sufficient
to
mimic
the
biological
patterning
process
again.
B
Sort
of
a
proof
of
concept
where
you
put
this
engineered
technique
into
a
biological
context,
and
you
see
if
it
produces
the
effect
that
you
want,
and
then
it
doesn't
necessarily
mean
that
you
have
the
mechanism
down.
It
just
means
that
sometimes
you
can
replicate
that
thing
that
you're
looking
to
seeking
to
reproduce.
B
B
So
this
is
a
an
example
of
some
of
these
different
mechanisms
and
how
they're
replicated
so
a
is,
let's
see,
is
lateral,
inhibition,
symmetry
breaking
mechanism
and
inner
ear
patterning.
So
this
is
an
example
here,
where
you
have
a
two-dimensional
cell
culture.
You
have
these
two
different
cell
States
and
you
have
pattern
formation.
B
B
This
is
so
sorting
B
where
you
could
sell
sorting
in
the
Inner
Cell
mass,
and
you
have
synthetic
sorting
on
the
right
which
is
in
vitro
again
not
in
the
context
of
a
developmental
body
and
you
get
the
same
sort
of
a
factory
an
outer
ring,
and
then
this
inner
mass
of
cells,
they're
different
states,
so
they're
sort
of
you
have
them
initially
in
this
sort
of
random
configuration,
and
then
they
form
this
pattern
and
foresee
you
have
stripe
patterning.
So
you
could
do
this
on
the
fish's
body
and
there's
in.
B
D
is
where
you
see
neural
crest,
migration,
and
then
this
is
the
in
in
vitro
version
where
you
can
actually
use
optogenetics
to
migrate.
The
cells
you
illuminate
the
the
cell,
the
area
and
the
cells
respond
to
that
elimination
of
stimulus.
You
can
also
do
polariferation
and
apoptosis.
This
is
the
classic
example:
a
digit
formation.
Where
cells
you
have
this
sort
of
mass
that
comes
out.
B
You
know
as
the
four
four
limb
that,
where
the
digits
are
going
to
form
the
hand
or
the
polar
or
whatever
it
is,
and
whatever
organism
you're
looking
at
and
then
you
have
this
pattern
cell
death,
where
it
forms
digits.
So
this
pad
becomes
a
set
of
digits
by
the
cell
intermediate
cells
dying
off
and
then
the
rest
of
the
cells
forming
these
digits,
and
you
can
do
a
similar
thing
in
vitro
where
you
can
use
optogenetic
apoptosis.
So
you
can
program
cell
death
using
an
optogenetic
stimulus,
and
so
these
are.
B
These
are
all
the
types
of
things
you
can
do
here.
You
can
actually
look
at
epithelial
folding
as
well.
This
is
an
example
in
the
tissue
where
you
have
an
all
tube
folding
and
you
can
use
optogenetics
again
to
fold
tissues
in
this
illuminated
area.
So
it's
a
stimulus
that
folds
the
tissues
in
this
simple
sheet
of
cells.
So
there
are
a
lot
of
ways
you
can
do
this
and,
of
course,
in
optogenetics
you
need
to
have
a
genetically
engineered,
Target
Gene
that
responds
to
the
optical
stimulus.
B
B
B
You
can
have
terminal
signaling
where
you
can
have
the
signal
when
you
can
remove
the
signal,
and
you
can
see
the
effect
in
this
guest
relation
of
this
embryo.
You
can
have
spatio
temporal
control
in
some
of
these
processes,
so
this
is
an
example
of
optogenetics
allowing
use,
control
and
space
and
of
shape
again
over
time
as
well,
because
you
do
this
at
a
certain
point
in
development.
B
So,
if
you're
building
a
developmental,
if
you're
engineering
a
developmental
model,
you
want
to
do
these
things
at
a
certain
time.
It's
not
that
you
just
want
to
do
these
things
whenever
when
now
you're
constructing
a
phenotype,
you
want
to
do
these
things
in
a
certain
order,
but
we
can
also
use
apical
constriction
to
alter
organoid
shape,
and
these
are
organoids
here,
where
you
have
some
putative
morphogenesis
going
on,
and
you
can
actually
refine
some
of
this
by
applying
some
of
these
techniques
to
an
organoid.
B
So
here's
an
optic
vesicle
organoid,
which
has
some
putative
morphogenesis
along
the
edges.
You
can
actually
shrink
different
parts
of
the
organoid
and
you
can
actually
shape
this
organoid
in
ways
that
you
couldn't
before
and
so
neuroactodermal
organoids.
You
have
a
similar
process
where
you
can
flatten
circular
morphology,
so
you
can
actually
form
shapes.
B
You
can
actually
encourage
morphogenesis
and
move
it
along.
So
this
is
an
interesting
review.
The
future
perspective
future
perspectives,
of
course,
that
you
have
two
things
here,
that
they
point
out:
combination
developmental
processes.
So
you
can
combine
these
different
things.
You
can,
for
example,
combine
reaction,
diffusion,
patterning
and
tissue
folding,
and
this
is
an
example
from
an
organoid
where
you're
able
to
take
a
circular
array
of
cells,
manipulate
some
of
the
cells
and
fold
them
out.
B
So
you
form
this
morphology
here
and
you
can
actually
sculpt
morphogenesis
out
of
these
very
simple
models,
but
you
can
also
study
complex
tissues,
so
you
can
actually
look
at
a
wild
type
shape
of
a
tissue.
You
can
perturb
the
shape
using
optogenetics
and
you
can
actually
ask
the
question:
what
are
the
feedbacks
here?
How
does
shape
get
informed
by
function
as
you're
eating
your
optigenetic
probe
you're?
Actually
probing
the
function
of
this
tissue.
So
what
is
the
effect
of
this
function
on
tissue?
And
you
can
do
this
in
a
controlled
manner?
B
A
B
B
To
the
top
here,
a
tug
of
war
between
filament,
treadmilling
and
myosin
induced
contractility
generates
actin
rings.
So
this
is
something
called
filament
treadmilling,
my
own
myosin
induced
contractility,
and
this
generates
active
rings.
So
this
paper,
in
most
of
your
periodic
cells
act
in
filaments,
assemble
into
a
shell
like
hectin
cortex,
which
is
a
sort
of
a
shallow
actin
under
the
plasma
membrane,
controlling
cellular
morphology,
mechanics
and
signaling.
So
this
is
something
that
happens
in
eukaryotic
cells.
B
You
get
this
actin
skeleton
forming
out
of
filaments
and
they
form
this
cortex,
which
is
a
Surface
outer
surface
or
outer
shell.
So
the
acting
cortex
is
highly
polymorphic.
Adopting
diverse
forms
such
as
the
ring-like
structure
is
found
in
autosomes
axonal
rings
and
immune
synapses
biophysical
principles
and
underly.
The
formation
of
actin
rings
and
cortices
remain
elusive.
B
B
We
found
that
actin
myosin
networks,
condense
into
clusters
at
low
treadmilling
rates
or
high
myosin
concentration,
but
form
ring-like
or
cortex-like
structures
or
high
treadmilling
rays
and
low
myosin
can
concentration.
So
this
is
like
something
that
they're
looking
at
these
different,
these
treadmilling
rates
and
myosin
concentration
and
they're.
Looking
at
the
sort
of
shapes
that
it's
it's,
the
shapes
that
are
being
made.
This
mechanism
is
supported
by
our
corroborating
experiments
on
live
T
cells,
which
exhibit
ring-like
actin
networks
upon
activation
by
stimul,
stimulatory
antibody,
upon
disruption
of
filament,
treadmilling
or
enhancement
of
myosin
activity.
B
The
pre-existing
actin
rings
are
disrupted
into
actin
clusters
or
collapse
towards
the
Network
Center
respectively.
So
this
is
the
structure
of
this
network
and
they're
looking
at
different
aspects
of
this
filament
treadmilling
and
myosin
activity,
and
so
these
rings
can
be
disrupted
in
different
ways.
They
can
collapse
towards
the
Network
Center.
B
B
B
They
use
this
platform
called
median,
which
is
mechanochemical
dynamics
of
active
Networks
and
experiments
on
live.
T
cells
they're
actually
doing
this
in
T
cells
with
their
modeling,
the
actin
Dynamics.
In
this
program
we
find
that
competition
between
actin,
filament,
treadmilling
and
myosin
contractility
determines
the
overall
Network
morphology.
B
A
B
Figure
one:
this
is
the
actin
and
mm2s
the
distribution
of
actin
rings
and
T
cells
in
the
simulation.
So
this
is
your
actin
ring.
This
is
your
mn2
and
this
is
the
simulation
setup.
So
you
have
this
nm2.
You
have
this
actin,
these
two
types
of
rings
and
you
put
them
into
the
simulation
around
the
edge
of
the
cell.
You
have
these
other
things
as
well,
and
then
you
have
this.
B
So
the
next
figure
shows
some
of
these
simulations.
They
have
their
minimal
model,
they're
able
to
measure
different
things
in
a
simulate
different
things
in
it,
they're
able
to
look
at
the
so
nm2
contractility
induces
geometric
collapse
of
treadmilling
actin
filaments,
normalized
medians
of
radio
filament
density
distribution,
different
treadmilling
rates
are
shown,
so
these
are
rates
of
where
they're
moving
around
and
they're
active,
and
they
show
that
in
some
cases
this
collapses
and
I
guess.
An
e
is
where
they
have
a
collapse
situation
actually
I.
Think
in.
B
B
A
B
B
So
this
shows
us
at
different
rates
some
of
the
simulation
results
here
and
so
from
that
they
can
actually
ask
some
questions,
one
under
the
energetic
origins
of
structural
polymorphism
in
these
active
Networks.
So
they
can
actually
look
at
so
now,
they're
exploring
the
chemical
and
mechanical
properties
of
actin
networks
of
various
treadmilling
rates.
We
found
the
numbers
of
effect
and
filaments
bound
linkers
and
bowel
Motors
are
linearly
constant
across
different
values
for
this
parameter.
B
Well,
distribution
of
diffusive
molecules
such
as
geactin
and
nucleators
also
did
not
show
spatial
location
being
uniformly
distributed
throughout
the
simulation
volume.
This
observation
suggests
that
ring-like
architectures
do
not
form
because
of
the
enrichment
of
soluble
constituent
molecules,
nailed
periphery.
So
this
is
where
ring-like
architectures
don't
form,
because
the
soluble
constituent
molecules
are
enriched
near
the
periphery.
In
other
words,
that
ring
doesn't
grow
in
volume
or
a
number
of
molecules.
B
So
it's
moving
around
this
ring
and
it's
not
necessarily
dependent
on
its
density.
The
lack
of
enrichment
of
soluble
molecules
in
the
periphery
suggested
the
possible
energetic
origin
of
these
structures.
We
must
examine
the
mechanical
energy
of
the
system,
which
primarily
arises
from
filament
bending
in
our
simulation
for
fixed
concentrations
of
nm2
and
cross-linker.
We
found
that
this
parameter
decreases
with
increasing
our
cement
at
TM.
In
addition,
emac
undergoes
a
sharp
reduction
where
our
sub
TM
reaches
the
critical
threshold,
a
few
of
actin
Rings
being
two
to
three
fold
lower
than
that
of
clusters.
A
B
B
This
is
no
inhibition
for
this
I
guess,
there's
a
a
chemical
treatment
that
they
do
here
where
it's
you
know,
nothing
weak
and
strong,
and
then
this
is
before
the
treatment
and
after
the
treatment.
So
this
is
before
the
treatment
in
800
seconds
for
the
vehicle
after
the
treatment
of
the
vehicle
which
doesn't
have
any
effect,
the
weak
inhibition,
which
is
a
low
concentration
of
lat
a
before
the
treatment
they
have
this
ring,
and
then
you
start
to
get
the
ring
starts
to
sort
of
collapse
down.
B
You
don't
see
very
much
of
an
effect,
though,
and
then
for
strong
inhibition.
This
ring
forms
because
the
ring
is
formed
as
it
is
in
the
other
examples,
and
then
the
ring
really
starts
to
break
down
after
it's
after
this
treatment
of
lad
a
so.
This
is
a
strong
dose
of
lead
a
and
it
forms
these
clusters
instead
of
this
ring,
so
the
ring
actually
can
collapse
down
in
the
Clusters,
and
so
that's
what
they're
talking
about
when
they
talk
about
clusters,
and
so
the
idea
is,
do
you
want?
B
You
know
it
in
the
ring
condition?
It
provides
this
structural
Integrity
in
the
cluster
condition
it
provides,
does
not
provide
that
so,
since
higher
our
median
indicates
localization
of
acting
filaments
at
the
network
periphery,
this
negative
correlation
indicates
that
configurations
of
the
lowest
mechanical
energy
are
those
of
the
ring-like
geometry.
So
that's
the
lowest
mechanical
energy
State
and
these
clusters
are
higher
level
mechanical
energy
states.
A
B
So
then,
this
detailed
mechanical
modeling
in
this
platform
shows
that
active
actin
networks
exhibit
a
striking
morphological
transition
upon
changes
in
the
filament
treadmilling
rate.
We
found
two
distinct
types
of
dynamic
structures
due
to
the
interplay
of
treadmilling
and
nm2
contractility
in
the
initially
disordered
Network
one
acting
clusters
form
and
slow
treadmilling
or
high
cnm2
networks.
So
this
is
the
treatment
of
this
drug
and
it
replicated
this
condition
and
two
ring-like
and
cortex,
like
structures
spontaneously
assembled
in
fast
treadmilling
on
the
cnm2
networks.
B
So
these
ring-like
structures
are
cortex
on
the
outside
of
the
cell
or
in
the
on.
The
outer
edge
of
the
cell
resulted
from
number
two.
This
geometric
transition
does
not
require
filament
tethering
to
the
boundary
where
a
spatially,
biased,
filament
assembly.
So
critically,
this
ring
isn't
the
product
of
self-organizing
I.
Guess
it's
self-organization,
but
it's
based
on
energetics,
not
on
any
sort
of
spatial
organization
or
some
sort
of
bias
in
the
in
its
position.
So
we're
we're
a
physical
mechanism
that
tethers
it
to
the
the
edge.
B
B
B
So
these
simulations,
although
they
agree
with
experiments
and
cells,
there
are
some
limitations.
So
one
limitation
is
that
they
did
not
explicitly
include
a
few
significant
properties
of
actin
networks.
So
they
can
only
include
some
of
the
mechanisms,
but
not
others,
and
that
may
have
an
effect
on
the
results
that
they
have.
B
However,
we
need
more
work
to
quantitatively
estimate
the
entropic
contribution
to
these
networks
and
their
self-organization,
so
we
don't
have
a
full
accounting
of
the
energetics
of
these
Networks.
So
this
is
something
that
needs
to
be
confirmed
with
future
empirical
studies,
and
so
that's
pretty
much
it.
This
paper
is
interesting
foray
into
the
world
of
actin
networks
and
minimal
models,
and
things
like
that,
so
the
second
paper
is
on
an
actin-based.
Vistoplastic
block
ensures
Progressive
body
access
elongation.
B
B
It
relies
on
the
interplay
between
intrinsic
forces
generated
by
molecular
Motors
extrinsic
forces
exerted
by
adjacent
cells,
mechanical
resistance
forces
due
to
tissue
elasticity
or
friction
understanding
how
mechanical
forces
influence
morphogenesis
that
the
cellular
molecular
level
remains
a
challenge.
Recent
work
has
outlined
how
small
incremental
steps
towards
Power,
Cell
autonomous
epithelial
shape
changes,
which
suggests
the
existence
of
specific
mechanisms
that
stabilize
Solitude
change
and
counteract
cell
elasticity.
So.
B
Pairs,
so
this
is
a
recent
work
here
that
they're
reviewing
and
they're
they're
engaging
with
how
is
outlined,
how
small
incremental
steps
Power,
meaning
that
they
influence
so
autonomous
epithelial
shape
change.
So
you
have
the
cell
autonomous
shape
change
mechanism.
This
thing
suggest
an
existence
of
specific
mechanisms
that
stabilize
cell
shapes
and
counteract
cell
elasticity.
So
this
is
this
lock
that
they're
talking
about
beyond
the
two-fold
stage
of
an
embryo,
embryonic
elongation
and
c
elegans,
and
so
we're
using
the
C
elegans
model
it's
dependent
on
both
muscle
activity
and
the
epidermis.
B
B
B
And
then,
as
you
get
cell
division,
of
course,
you
start
to
get
elongation
along
these
two
axes.
P1
is
here,
A
B
is
here
and
we
know
the
sort
of
the
sub
lineages.
So
these
these
cells
form
cell
lineage
and
a
b
forms
of
someone.
Hp
ones
were
there's
a
cell
lineage
and
they
tend
to
be
in
these
two
sides
of
the
embryo,
the
anterior
and
the
posterior,
and
and
then
there's
this
elongation
that
happens
within
the
embryo
to
pull
it
apart,
so
that
it's
elongated
So.
B
Eventually
you
get
like
a
comma
stage,
whether
something
like
this
in
a
pretzel
stage,
where
the
worm
is
kind
of
the
pretzel
and
then
finally,
it's
like
this
sort
of
spiral
shaped
morphology
that
will
hatch
into
a
larva
and
The.
Larva,
of
course,
is
a
fully
elongated
and
running
around
and
behaving
with
muscles
and
everything
the
muscles
start
to
form
and
the
epidermis
star
perform
around,
maybe
in
the
comma
stage.
But
what
they're
arguing
is
that
this
polarity
starts
early
and
continues
out,
and
these
mechanisms
pull
the
embryo
in
these
different
directions.
B
B
B
We
search
for
factors
of
genetically
or
molecularly,
interacting
with
the
p21
activating
kinase
homolog
pac-1
and
acting
in
this
pathway,
thereby
identifying
the
alpha
Spectrum
spc-1-
and
this
is
a
lot
of
molecular
biology
here.
So
combined
absence
of
pac-1
and
spc-1
induced
complete
axis
retraction
owing
to
defective
epidermal
actin
stress,
fiber
modeling
predicts
that
a
mechanical
vistaplastic
deformation
process
can
account
for
embryo
shape.
B
Stabilization
molecular
analysis
suggests
that
cellular
basis
for
viscoplasticity
originates
from
Progressive
shortening
of
epidermal
microfilaments
that
are
induced
by
muscle,
contractions,
relayed
by
actin,
severing
proteins
and
from
form
and
homology
to
domain
containing
protein
1
or
fhod1
Mormon
bundling
our
work.
This
identifies
an
essential
molecular
lock,
acting
in
a
developmental
ratchet-like
process.
So
this
is
a
sort
of
a
ratchet
and
a
lock
model,
which
is,
of
course
where,
if
you
have
a
ratchet,
you
move
it
forward
and
it
moves
your
bolt.
B
Maybe
if
you
have
a
ratchet
on
like
a
bolt
or
something
like
that,
it
moves
it
in
One,
Direction
and
then
there's
a
lock
so
that
it
doesn't
move
back
in
the
other
direction.
When
you
move
the
ratchet
back
to
make
another
turn
so
ratchets
are
kind
of
they
exert
forces
in
One
Direction
and
then
in
order
to
get
the
ratchet
back
to
its
starting
point
to
make
the
next
tightening
you
need
to
move
it
back,
but
of
course
it
needs
to
lock
in
place
in
order
to
do
that.
B
B
So
there's
this
retraction
phenotype.
So
they
have.
We
conclude
the
mechanical
input
provided
by
muscles
to
the
epidermis
induces
the
retraction
phenotype
observed
in
spc-1
PAC
pak1
double
mutants,
so
you
can
actually
have
this.
This
mechanism
can
vary
based
on
the
mutant
that
you're
looking
at
so
when
C
elegans
we
had
to
find
mutants.
B
We
can
actually
look
at
some
of
these
things
with
spc1
and
pack,
one
mutants.
We
can
look
at
how
these
embryos
so
in
these
double
mu
or
actually
yes,
and
these
double
mutants
embryos
could
reach
about
65
microns
at
a
slow
rate,
but
then
failed
to
maintain
their
shape
and
retract
it
back
to
about
50
microns,
so
they're,
looking
at
the
length
of
the
embryo
this
this
elongation
here.
B
Because
muscles
are
tightly
mechanically
coupled
to
the
epidermis
they're
epidermal
hemidemisomes,
their
contractions
also
displace
the
epidermis.
So
this
is
what
allows
for
this.
This
shape,
change,
and
so
there's
this
stretching
that
occurs,
and
this
ratchet
ratcheting
mechanism
is
to
maintain
the
ability
to
maintain
this
stretching
outward,
and
so
these
in
these
double
mutants.
You
don't
have
this
lock
on
the
ratchet.
B
So
a
simple
hypothesis
would
be
that
some
mechanism-
stabilizes
the
transient
cell,
shapes
and
display
muscle
activity,
for
example,
during
drosophila,
gastrulation
and
germ
band
extension,
which
is
in
drosophil
embryo,
is
a
similar
process
to
this.
An
infectile
embryos
have
a
longer
germ
than
C
elegans
by
far,
and
so
they
do
this.
This
sort
of
stretching
process
all
right,
ectomyosin,
pulsatile
flows,
are
thought
to
be
progressively
modified,
Junctions
or
thought
to
progressively
modify
Junctions
tongue
covers
such
a
mechanism
and
C
elegans.
B
A
B
This
is
where
it's
stretched
out,
and
this
is
an
example
here
of
in
I
guess
in
the
comma
stage
here,
where
you
get
the
stretching
this
elongation
and
you
get
these
muscle
contractions
that
they're
measuring
here
all
right.
So
in
this
figure
we
can
see
a
model
in
F
we
have
a
cellular
model
of
embryo
elongation
based
on
volume
conservation
in
our
librios.
B
So
this
is
a
case
where
you're
looking
at
volume
conservation,
you're
looking
at
the
activity
of
pac-1
and
spc1,
this
actin
binding
protein,
so
during
contraction,
for
example,
in
the
wild
type.
You
get
this
activity
here
and
contractions,
so
you
can
see
them
being
pushed
together
and
relaxation
they're
not
being
pushed
together,
and
so
this
is
normal
remodeling
in
effective
remodeling
in
these
double
mutants.
B
You
have
this
action
of
contraction
and
then
this
action
of
relaxation
and
during
relaxation
these
filaments
break
down
here
so
that
there
there's
this
defective
remodeling
that
occurs
and
doesn't
really
stay
in
place.
These
these,
in
this
case
relaxation
means
that
they
lock
in
place
in
the
defective
Remodeling
and
relaxation
everything
actually
relaxes
and
there's
no
tension
informed
to
maintain
this
change.
B
B
The
spring
is
here,
and
this
kind
of
shows
the
forces
of
the
epidermis
and
muscles
acting
on
this
plastic
spring
and
then
that's
that's
how
they
model
the
spring
and
dash
using
a
springer,
Dash
pod
model,
and
then
this
D
here
shows
some
of
these
mutants
and
the
limo
lima
bean
stage,
the
lima
bean
plus
130
minutes
and
the
Lima
beam
plus
220
minutes.
And
so,
as
you
can
see,
this
is
what
happens
in
these
different
mutants
diamonds.
F
hard
one
spc1
mutant,
it's
actually
treated
with
SBC
one
is
treated
with
rnai.
B
So
it's
interfering
with
that
Gene,
the
expression
of
that
Gene,
and
you
can
see
that
it
has
this
sort
of
it
achieves
a
comma
period
at
a
plus
130
minutes,
because
and
then
it
goes
back
to
this
shape
here
and
this
one
has
pc1
with
rnai.
It
achieves
the
comma,
but
it
kind
of
maintains
it
in
this
case.
It
sort
of
maintains
it
a
little
bit
in
this
case.
It
may
doesn't
maintain
it
very
well,
and
so
you
can
see
that
how
this
interferes
in
different
ways.
Sometimes
it
maintains
the
shape.
Sometimes
it
doesn't.
A
B
Okay,
and
so
in
conclusion,
you
know
they're
they're
kind
of
investigating
the
different
factors
that
might
involve
a
breakdown
of
this
ratchet
mechanism,
so
some
of
it
involves
the
failure
of
actin
bundling
and
things
like
that.
There's
also
activation
of
the
muscle
induced
mechanic.
Transduction
pathway
could
be
affected.
B
We
conclude
that
actin
filament,
severing
initiated
by
muscle
contractions
following
by
fhod
one
dependent,
bundling
or
capping,
represents
a
ratchet-like
mechanism
providing
a
molecular
basis
for
vistoplasticity,
and
so
we
don't
know
all
the
factors.
But
we
know
from
these
experiments
that
there
are
some
mutants
that
don't
exhibit
this
sort
of
ratcheting
process
in
summer
results
identify
several
proteins
that
are
involved
in
stabilizing
cell
shapes
and
assistance
subjected
to
repeated
external
mechanical
stress.
We
propose
that
the
progressive
shortening
of
actin
filaments
under
the
control
of
these
factors-
adiates
a
cellular
vistaplastic
processor,
promotes
axis
elongation.
B
A
similar
vistoplastic
process
might
operate.
Invertebrate
tissues
comprising
an
epithelial
layer
surrounded
by
a
contractile
layer,
so
invertebral
tissues.
There
might
be
some
sort
of
process
where
you
have
an
epithelial
where
and
then
in
in
the
middle
and
then
a
contractile
layer
on
the
top
something
like
internal
organs
in
humans,
so
this
could
actually
be
like
a
composite
version
of
this,
which
we
see
in
C
elegans,
where
there's
different
types
of
contraction
and
growth.
So
such
a
process
might
therefore
also
influence
the
medicine
metastatic
properties
of
tumor
cells
juxtaposed
with
contractile
cells.
B
B
And
so
this
would
be
like
a
something
where
you
might
be
able
to
control
this
epithelial
Mass
inside
of
it
squeeze
it
or
you
know,
control
the
shape
of
it
in
some
way,
so
they're
suggesting
that
this
is
something
that
happens
in
metastatic
cancers
or
there's
as
where
the
tumor
metastasizes,
so
that
that's
how
they're
thinking
about
that
all
I
wanted
to
talk
about
in
terms
of
those
papers
help
you
learn
something
yeah.
So.
B
B
A
B
Always
a
thing
to
keep
in
mind,
though,
with
like
engineering
and
some
sciences
that
they
make
assumptions
in
the
paper,
and
then
they
don't
tell
you
what
those
assumptions
are.
So
you
always
have
to
think
about.
Like
you
know,
is
this:
is
this
the
only
way
it
can
be
done
or
are
they
just
kind
of
throwing
in
something
as
an
example?
Sometimes
people
just
don't
make
that
clear?
It's
like
we'll
just
assume
this,
and
then
you.