►
From YouTube: DevoWorm (2021, Meeting 38): Hacktoberfest, Planarian Body Plan Network, Regulatory Networks.
Description
Overview of progress on Hacktoberfest. Regeneration and the Planarian body plan network, theoretical types of developmental networks. NASA's open-science initiative, professional development discussion, and Virtual Developmental Biology. Eric Davidson's Regulatory Genome and its computational implementations (towards a computational version of the GAL4/UAS system). Attendees: Susan Crawford-Young, Jesse Parent, and Bradly Alicea.
A
C
A
A
D
C
A
C
C
C
C
So
this
is,
you
know
one
of
the
questions
that
they've
had
is
like.
Since
the
cells
are
toady
potent
and
they
generate
new
animals,
do
they
behave
like
they're
ancestral
animals,
and
there
have
been
some
studies
that
have
been
done
where
they've
shown
that
there's
some
sort
of
memory
that's
carried
over
between
the
cells.
In
other
words,
if
you
train
a
planarian
on
some
task
like
some
simple
behavioral
task-
and
this
is
what
they're
referring
to
here-
you
know
that
that
planario
can
learn
things
like
the
simple
associative
learning.
C
But
then
you
can
take
that
planaria
and
break
it
up
into
its
constituent
cells
and
put
those
cells
down
in
media,
and
then
you
can
get
the
you
know.
You
can
get
whole
organisms
to
regenerate
from
those
and
they'll
remember
what
they
learned
when
they
were
back
when
they
were
like
a
singular
planaria.
C
So
it's
really
interesting-
and
this
looks
like
some
of
the
stuff
they're
doing
with
looking
at
the
different
types
of
behavior.
So
this
is
actually
interesting.
What
is
this
graphical
abstract?
So
this
is
the
turn
elongated
contract,
so
they
have
the
different
parts
of
the
body
here
and
what
is
this?
C
C
C
We
found
that
in
response
to
mechanical
and
near
uv
stimulation,
so
this
is
with
like
simple
touch
or
simple
light
stimuli
in
the
uv
band
or
near
uv
band,
head
stimulation
precedes
turning
tail
stimulation,
precedes
contraction
and
trunk
stimulation
produces
mid-body
elongation,
so
they're
different
three
different
types
of
behaviors.
When.
A
C
A
C
Yeah
so
then
they're
able
to
do
this
in
a
certain
species
of
planarian
when
cut
into
two
or
three
pieces,
the
anterior
end
of
each
headless
piece,
so
they're
cutting
off
the
headlight
they're
cutting
off
the
head
and
they
have
two
or
three
pieces,
so
I
guess
they're
cutting
it
at
the
turn.
Elongation
and
contact
boundaries
shown
down
here.
C
B
C
Behavior
right
to
the
turn
behavior,
so
it
responded
as
if
I
thought
it
was
a
head.
In
addition,
posterior
ends
of
the
head
and
mid-body
pieces
sometimes
produce
contraction
instead
of
elongation.
So
in
other
words,
this
is
these
posterior
ends
here
would
behave
as
if
it's
like
the
tail.
So
basically,
if
I
cut
this
part
out,
this
red
part
it'll
start
to
behave
like
the
entire
worm.
C
With
this
end,
turning
and
this
end
contracting
and
the
middle
elongating
and
then,
if
it
matures
it
will
take
those
you
know,
it'll
be
assigned
those
different
behaviors
accordingly.
So
that's
interesting,
you
know
it's
not
just
that
the
cells
are
totally
potent,
it
seems,
like
the
entire
organism,
is
sort
of.
Has
this.
C
C
Sort
of
chemical
signaling
mechanisms
as
well-
I
don't
think
people
really
know
exactly
what's
going
on,
but
this
is
a.
This
is
a
fun
read
yeah.
The
observations
show
that
each
mid-body
region
reads
the
global
state
of
the
organism
and
adapts
its
response
to
incoming
signals
for
the
remaining
tissue.
C
C
All
right,
yeah
we're
just
talking
about
the
great
world
of
planarians,
our
flatworms,
and
they
found
a
paper
on
where
susan
found
a
paper
on
regeneration,
and
I
don't
know
if
you've
read
some
of
the
stuff
on
that,
but
that's
pretty
fascinating,
especially,
I
know
mike
levin.
If
you
know
some
of
his
work
on
planarians
has
done
a
lot
of
work
on
regenerating
planarians
and
how
they
can
sort
of
figure
out
what
the
morphology
should
be
from.
C
C
The
open
worm
slack
in
the
diva
worm
channel
okay.
So
I
wanted
to
talk
a
little
bit
about
some
of
the
things
going
on
this
month.
We
have
our
hacktoberfest
is
ongoing,
we're
you
know
we're
not
getting
a
lot
of
participation,
but
we
are
getting
usual.
Faroon
and
esme
were
featured
at
last
week's
meeting
and
they
were
talking
about
some
of
the
stuff
they're
doing
with
basil
area.
C
And
then
you
can
use
that
to
you
know,
find
you
know,
find
a
lot
of
different
types
of
like
what
they're
trying
to
do
is
they're,
trying
to
train
they're,
trying
to
create
a
skeleton
of
rectangles
out
of
the
cells,
and
then
they
want
to
use
that
as
a
as
a
way
to
identify
cells
over
time.
So
they
want
to
break.
They.
Take
these
movies
of
motion
in
the
microscopy
images
of
motion.
C
So
you
have
this
series
of
images
where
you
know
the
cells
are
moving
back
and
forth
in
the
frame
and
they
want
to
find
the
cells
more
effectively
in
that
type
of
setting
where
you
have
to
identify
them.
Maybe
if
they're
cut
off
by
half,
if
they're
off
the
frame
or
if
they're
you
know
kind
of
moving
in
a
weird
position,
if
they're
in
a
weird
position
as
they
move,
you
know,
how
do
you
find
those
and
so
they're
working
on
that?
C
Hopefully
they'll
be
able
to
show
some
results
in
a
few
weeks,
but
and
for
those
of
you
who
want
to
participate,
I'm
going
to
share
my
screen
so
those
who
want
to
participate
in
oktoberfest.
I
think
this
is
it
here
okay,
so
this
is
the
digital
basilaria
repository
and
we
have
the
hacktoberfest
badge.
C
C
So
you
might
want
to
inquire.
I
don't
know
we
have
some
issues
here.
That
might
need
to
be
addressed,
but
they're
kind
of
older.
So
I
don't
know
if
they're
going
to
be
of
use,
then
we
also
have
diva
learn.
So
divalern
is
the
of
course.
The
platform
we've
been
working
on
for
the
last
two
years
and
we
have
the
diva
learn
software,
which
has
another
release
coming.
C
What's
there,
I'm
not
really
sure
what
there
is
right
now,
that's
outstanding
that
can
be
achieved
in
a
short
period
of
time,
we're
working
on
a
paper
for
this,
but
this
is
something
that
I
think
it's
going
to
be
amongst
you
know,
maybe
three
or
four
of
us
to
do
that,
and
so,
but
in
diva
learning
the
organization
we
have
other
things
going
on.
We
have
the
data
science
demos,
which
are
always
a
good.
C
I
mean
we've
had
probably
about
eight
people,
contribute
data
science,
demos
here
and
if
you
have
a
topic
that
you
want
to
cover
as
a
notebook
or
as
some
sort
of
tutorial
that
you
can
put
on
to
github
and
have
the
notes
available
for
people
to
follow
along,
then
that's
something.
Maybe
you
can
contribute,
so
that's
diva,
learn
and
then
I
wanted
to
kind
of
cover.
C
Okay.
So
I
think
that's
it
for
this.
That's
digital
basil
area.
We
have
data
science,
demos
and
then
we
also
have
the
group
meetings
repository
and
that's
not
as
straightforward,
because
I
don't
think
we
have
well.
We
have
a
lot
of
issues,
but
those
are
mainly
on
our
project
board.
So
you
can
contribute
by
going
to
the
group
meetings,
you
can
go
to
the
project
board
here,
which
is
this
group
meetings
projects
one
and
you
can.
You
know,
look
over
some
of
the
many
issues
that
we
have
open
here.
C
C
So
things
that
are
on
hold
you
know
some
of
these
are
just
kind
of
things
that
we're
talking
to
other
people
with
things
that
people
are
already
get
engaged
with,
but
they
don't
have
time
to
finish
now
so
and
then
to
do
so.
There
are
a
lot
of
things
we
can.
We
can
follow
up
on,
as
virtual
developmental
worlds
is
is
interesting.
C
D
C
This
one
is,
I
think
it's
not
loading.
For
me,
this
is
the
one
where
we're
looking
to
build
like
a
virtual
environment
or
a
3d
set
of
3d
environments
for
developmental
processes.
You
know
maybe,
like
we
were
talking
at
one
time
about
building.
You
know
3d
models
of
embryos
or
putting
something
into
the
blender
container
for
open
worm.
C
But
I
think
this
is
a
more
general
theme
where
you
know
we're
trying
to
imagine
sort
of
like
three-dimensional
worlds
for
developmental
processes,
so
that
could
be
like
you
know,
3d
modeling
in
blender,
or
it
could
be
some
sort
of
like
you
know,
vr
experience
or
it
could
be.
You
know
just
like
a
three-dimensional
model
or
a
higher
dimensional
model
of
development.
I
mean.
D
C
C
You
know
that
that
might
be
an
interesting
thing
to
do
like
have
a
almost
like
a
you
know,
a
fitness
based
thing
where
you
have
like
an
embryo
in
a
world,
and
then
it
has
like
physical,
physical
environment,
where
it's
not
really
exploring
the
environment.
So
much,
but
it's
you
know
it's
exposed
to
an
environment
with
like
light
with
surface
properties.
With
you
know,
heat
then
they're
also,
you
know
where
we
could
have
like
little
agents.
C
That
are,
you
know,
very,
like
sort
of
hatching
out
of
their
egg
and
they
have
like
no
experience
in
the
world
and
they
have
to
sort
of
find
their
way
out
of
an
egg.
You
know
I'm
just
thinking
here
about
like
how
you
know
if,
if
an
organism
comes,
sometimes
organisms
are
pretty
well
developed
coming
out
of
their
egg,
but
sometimes
they're.
Not
so
you
know
what
are
they
encountering
as
they
first
enter
the
world?
C
D
C
D
Genetic
algorithm
or
something
yeah,
I
don't
know
if
they
reached
out
to
you
yet,
but
this
is
more
for
the
other
group
monkey
grover
dm
me
yesterday
about
stuff,
okay
guys
he
asked
me
to
like
he
asked
me
about
one
of
the
things
that
I
said
to
mention.
I
I
tend
to
follow
up
with
you.
I
don't
know
if
he
did,
I
think
he's
gonna
try
to
do
it
in
slack.
Okay,
so
just
a
note
that
he
was
going
to
talk
about
like
one
of
the
switching
papers
or
something
I
think
so.
Okay.
C
C
C
C
Well,
they
kind
of
do
have
incentive
to
do
this,
but
because
you
know
there
are
a
lot
of
private
agencies
now
that
are
doing
space
exploration.
So
this
is
you
know
this
is
their
open
science
open
source
science
initiative,
so
nasa
is
making
a
long-term
commitment
to
building
an
inclusive,
open
science
community
over
the
next
decade.
C
This
is
includes
the
open
sharing
of
software
data
and
knowledge,
algorithms
papers,
documents
and
solar
information
as
early
as
possible
in
the
scientific
process.
The
principles
of
open
source
science
are
to
make
publicly
funded
scientific
research,
transparent,
inclusive,
accessible
and
reproducible.
C
A
C
Yeah,
I
don't
know
if,
like
they
have
a,
I
don't
know
if
that
includes
like
you
know
how
you're
putting
experiments
like
in
a
on
the
space
station
or
something,
but
that
would
be
interesting
yeah.
Well,
if
that's
what
it
is
yeah
yeah
I
mean
I
don't
yeah,
maybe
they'll
they'll
start
putting
those
like
if
they
do
some
experiments
in
space
they'll
be
opening
up
those
data.
So,
like
they've
done
this
with
a
lot
of
different
organisms
where
they've
you
know,
because
people
have
similar
similar
questions
to
that.
C
A
Yeah,
well,
this
one's
quite
quite
a
clever
idea,
because
if
you
disrupt
the
microtubules,
they
don't
develop
what
the
holes
or
the
pores
in
their
frustules,
they
it
becomes
more
of
a
clumped
process,
so
anyways
it
he
he
put
this
together.
He
applied
to
the
canadian
nasa
branch
and
then
they
said,
oh
well,
we're
not
doing
biological
experiments
from
canada.
This
is
like
yeah
yeah,
with
a
lot
of
bother
for
nothing.
C
A
C
C
C
We
have
these
molecular
level
simulations
of
diatoms
and
movement
simulations.
That's
something
that
I
think
our
group
that's
doing.
The
digital
digital
basilari
might
be
able
to
address
with
some
of
the
data
they're
collecting,
so
once
they,
you
know
perfect
their
algorithm
and
they
kind
of
go
through
some
of
the
data
that
they
have.
You
know
we'll
have
some
numeric
data
that
we
can
then
look
at
and
analyze,
so
we
might
actually
be
able
to
address
these
fairly
soon.
C
This
quantitative
comparison
of
our
key
and
shape
droplets-
that's
I
haven't
heard
about
that
in
a
while
I'll
have
to
check
back
on
that.
That's
I
think
my
knock
was
working
on
that,
but
and
then
this
neuro
match
submission.
So
I
mostly,
I
think,
most
of
the
activity
on
that's
going
to
be
from
what
the
other
group
that
I
work
with.
C
I
was
also
thinking
of
putting
on
like
a
parallel
session
with
either
diva
warm
or
open
worm
kind
of
topics
over
you
know
during
neuromatch,
so
it
wouldn't
necessarily
be
sponsored
by
narrow
match,
but
it
would
be
concurrent
with
neuromatch
and
I
don't
know
maybe
we
can.
You
know
get
the
word
out
that
this
is
happening.
C
I
I
don't
know
exactly
what
it
would
look
like.
I've
been
thinking
about
it
for
a
while,
like
what
would
this
look
like?
Would
this
be
something
that
we
would
have
like
a
you
know,
a
call
for,
and
then
people
say.
Oh
yeah,
I
like
development,
and
I
like
the
brain
and
let's
do
something
around
that,
and
I
don't
know
if
it
should
be
like.
C
I
don't
think
it
would
be
like
a
hackathon,
but
it
could
be
like
something
like
that
where
people
just
come
and
try
out
ideas
or
talk
about
ideas,
or
it
could
be
something
like
just
a
presentation
of
some
of
the
things
that
we've
been
doing,
I'm
never
always
sure
how
to
do
that.
I
don't
really
want
it
to
be
like
an
advertisement
for,
like
you
know,
it's
kind
of
hard
to
you
know.
People
get
disappointed
in
that.
C
So
yeah.
That's
what
we're
working
on
in
terms
of
submissions
kind
of
like
the
same
as
it's
been,
it's
been
the
stasis.
I
haven't
seen
too
many
opportunities
come
forth
for
other
types
of
things
again
the
nurips,
if
you're
interested
in
europe's
that
is
coming
up
in
late,
or
I
think
the
conference
starts
in
early
december.
C
D
D
D
In
a
in
an
ideal
sense,
it
might
be
sort
of
like
a
way
to
discuss
or
like
relevant
a
relevant
discussion
of
topics,
but
also
things
that
we
kind
of
dovetail
into
actual
projects.
We
want
to
do
or
would
be
doing
in
like
going
into
you
know
the
next
year
like
what's
going
to
be
happening
in
in
january
february
march.
It
was
kind
of
like
the
pre,
the
pre
google
summer
code
period,
oh
yeah,
something
to
think
about.
C
Yeah,
like
I
usually
do
well,
I
tried
to
do
yearly
recap
at
the
end
of
the
year,
but
that's
pretty
much
close
to
the
end
of
the
year,
anyways,
so
yeah
going
into
yeah.
I
guess
that
could
be,
and
then
the
question
is
is
like:
how
do
you
get
people
to
notice
it
or
is
it
something
that
would
be
like
concurrent
with
neuromatch
or
is
something
maybe
we
pre-record
and
put
up
dirt?
You
know,
I,
I
don't
know
the
logistics
of
that.
What
would
be
best.
C
Yeah,
I
think
that
yeah
and
then
that
might
be
a
good
idea
to
have
an
open
science
component,
just
kind
of
like
going
over
some
of
the
way.
Maybe
the
way
we
do
it
here
and
then
you
know
maybe
a
more
general
way
of
doing
it
like
getting
people
familiar,
maybe
with
the
incf
resources
and
some
of
the
other
resources
that
are
like
more
neuro-oriented.
C
You
know
some
of
the
other
things
that,
like
with
open
worm,
so
open
worm.
I
I
mentioned
this.
I
proposed
this
during
the
open
worm
week
annual
meeting
and
they
were
like
you
know:
oh
yeah
yeah,
it's
okay,
you
know,
I
don't
know
how
much
we're
gonna
get
participation
out
of
the
different
groups.
I
don't
think
it's
gonna
be
like
this.
C
C
C
You
know,
like
you,
know,
looking
at
simulating
ion
channels
simulating
worm
movement
in
in
physical
environments,
so
people
have
done
that
at
a
very
high
level
and
there
are
a
lot
of
participating
labs
in
conjunction
with
open
worms.
So
a
lot
of
those
are
in
europe
and
you
know
a
lot
of
c
elegans
labs
are
always
interested
in
this.
So
it's
it's
like
I
don't
know.
Maybe
I
think
it's
underutilized
the
component,
where
you
say
okay,
I
want
to
get
involved
in
open
worm.
It
might
actually
be
relevant
to
my
work.
C
A
D
For
both
yeah
okay
and
one
more
tangential
idea
that
just
came
up
or
more
so
a
question
and
that's
kind
of
interesting,
because
it's
about
the
the
context
of
my
question
is
obviously
I'm
applying
to
graduate
school
and
updating
my
cv.
And
I'm
using
I'm
rearranging
a
bunch
of
subcategories
and
cv
and,
like
you
know,
the
sort
of
nebulous
questions
about
like
service
and
and
and
how
you're,
how
you
show
off
different
things
that
you're
doing
and
I'm
in
the
process
of
doing
that.
And.
D
But
I
also
feel
like
I'm
hearing,
I
mean
I've
just
thought
of
it
not
more
looked
into
it,
but
I
also
feel
like
I
and
some
of
the
feedback
that
I've
gotten
and
also
like
just
to
kind
of
say
it
here.
I'm
you
know,
I'm
for
those
who
don't
know
me
about
it.
I'm
part
of
a
fellowship
program
and
part
of
the
fellowship
program
is,
is
doing
these
mentoring
sessions
and
discussion
intersessions
with
our
the
cohort
of
the
people
in
the
fellowship,
and
one
thing
that
came
up
in
one
of
the
sessions
was.
D
You
know,
like
it's
great,
to
show
specifically
your
your
your
diversity
inclusivity
activities
in
your
applications
in
your
cbu
or,
however,
you're
addressing
that
for
graduate
school,
because
that's
a
very
important
thing-
and
there
was
talk
about
that
and
some
things
that
I
do
that
and
I
thought
well,
we
didn't
what
we
didn't
talk
about
yet
and
I'll
bring
up
next
time
there.
But
it's
the
open
science
component
like
like
how
do
you
you
know,
address
that
or
label
that?
Even
so,
I'm
just
thinking
about
that
out
loud
right
now,.
C
C
You
you
might
want
to
have
an
open
science
category
just
on
you
can
do
this
on
your
cv
to
put
open
science
activities,
and
then
you
put
those
in-
and
I
know
a
lot
of
the
things
you've
done
like
you
know
like
with
neuro
match
and
other
things
with
the
orthogonal
web
are
kind
of
non-traditional
like
people
generally,
don't
do
them.
You
know
it's
not
something
that
everyone
gets
exposure
to.
So
that
would
be
good
under
open
science
activities,
because
it's
not
really
service
yeah,
but
it's
not
really
anything
else.
D
I'm
I'm
almost
for
me
personally,
because
I
I've
I
I
for
a
long
time,
misunderstood
the
service
category
and
then
I
talked
about
it
with
them
with
the
mentor
and
it
was
sort
of
like
you
know
like
what
I
what
I
did
for
hosting
the
cogside
panel,
like
that
kind
of
stuff
or
being
a
moderate,
the
cockside
discussion
group
and
then
a
bunch
of
the
like
I've
done
a
lot.
I've
done
panels
like
with
your
celebration
of
computing.
I've
been
on
my
alumni
stuff.
I've
been
a
moderator
for
something
with
you
and
elsewhere.
D
I've
been
moderators
like
those
kind
of
all,
okay
that
makes
sense
and
like
even
even
specifying
what
those
are
and
like
getting
used
to
categories
of
colors.
But
then
I
think
I'm
thinking
like
I'm
almost
I'm
I'm
just
not
considering
an
open
science
thing
and
also,
potentially,
I
might
just
be
open
science
and
or
like
kind
of
a
mentoring
or
not.
I
really
want
to
say
professional
development,
but
that's
sort
of
adjacent
to
what
we've
been
saying
in
the
other
group.
D
So
there's
there's
sort
of
it's
a
mix
between
those
are
things
that
I'm
doing,
and
also
I
have
a
lot
of
stuff
that
I'm
trying
to
cut
down,
because
I
don't
want
my
stuff
to
be
too
long,
even
though
cvs
are
supposed
to
be
long
like,
I
don't
want
to
bury
the
leak
about
anything,
so
it
I'll
show
you
I'll
show
you
more
later.
This
is
kind
of
just
things,
I'm
thinking
about
in
a
moment
right
now,
yeah,
but
I
appreciate
those
discussions.
Yeah.
C
C
You
know
things
like
that
and
then
you
know
you
can
you
can
have
whatever
categories
you
want.
Really
I
mean
I
have.
I
don't
know.
I
probably
share
some
of
my
cvs.
I
have
like
a
main
cv
which
is
really
long,
and
then
I
have
some
alternate
cvs
which
are
shorter,
which
focus
like
I
have
one.
That's
open
science,
one!
That's
like
just
selected
publications.
C
So
there
are
a
lot
of
different
like
ways
you
can
tweak
it.
Just
for
the
like,
because
you'll
find
that
like
when
you
write
a
grant.
For
example,
you
have
to
have
a
very
short
report
of
what
you're
doing
like
just
be
like
two
or
three
pages,
so
that
and
but
then
also
you
want
to
have
a
longer
one.
So
you
can
show
people
what
you've
done
so
having
like
this,
you
know
maybe
different
versions
or
it's
a
module
where
you
can
take
things
and
cut
them
out
or
put
them
in
depending
on
what
you.
D
C
D
I'm
not
really
sure
what
to
make
of
it.
I
like
basically
it's
an
attempt
at
combining
your
orchid
stuff
and
like
automatically
pulling
from
that
sort
of
and
some
other
adjacent
things
which
makes
sense,
but
also
like.
I
wonder
if
I
kind
of
haven't
really
invested
that
much
into
orchid.
But
I
wonder
if,
for
funding
purposes
and
everything
else
at
this
point,
it
might
be
better
to
put
things
on
there,
because
that's
like
a
very
there's
a
tied
to
like
your
nsf
stuff
right,
like
it's
very
so.
C
Yeah
yeah,
that's
just
usually
for
well,
you
can
put
publications
and
you
can
build
stuff
on
there
yeah.
I
would
do
that.
Put
everything
in
I
you
know.
There's
like
there
was
a
period
where
I
was
putting
them
into
all
these
different
tools
like
work
it
and
all
these.
So
sometimes
you
have
to
be
redundant
and
put
things
in
different
places.
I
definitely
work
update
my
orchid
profile,
though,
because
they'll
use
that
for
like
nsf.
A
C
D
D
How
much
is
is
the
juice
worth
the
squeeze
kind
of
a
deal
for
it,
because
it's
sort
of
save
your
own
thing,
but
also
I
have
to
I
have
to
put-
I
don't,
have
anything
in
orchid
right
now
and
like
some
things,
I
need
to
have
it's
it's.
It's
a
really.
It's
a
nice
conversation
and
we
don't
talk
about
it
anymore
here,
but
like
just
just
a
lot
of
it's
a
very
interesting
time
for
me,
I
feel
like
between
now
and
the
end
of
the
year.
B
C
So,
actually,
I
wanted
to
share
some
more
a
couple
more
things
here
and
then
we'll
move
to
papers.
This
is
the
I
I
want
to
go
over
this,
because
this
is
maybe
tied
into
you
know
what
we
might
be
doing
in
the
future.
This
is
something
actually
that
I've
I've,
so
at
networks
2021
and
then
I
did
another
presentation
on
this
at
another
workshop.
C
There's
this
idea
of
embryo
in
developmental
networks
and
how
they're
sort
of
divergent
from
one
another
in
development.
So
you
have
like
a
bunch
of
cells
and
an
embryo,
and
then
you
start
to
get
neurons
that
differentiate
and
then
those
neurons
will
wire
into
a
neural
network
and
then
the
embryo
cells
will
differentiate
into
different
tissues.
C
Sort
of
as
a
group
of
a
group
of
cells-
so
you
know
when
you
get
a
maybe
some
sort
of
critical
mass
of
neurons,
like
maybe
10
they
start
to
wire
into
this
network
and
when
they
start
to
wire
into
the
network,
they
have
their
own
behaviors.
They
interact
with
one
another.
They
affect
one
another
through
signaling.
C
Thing
when
a
tissue
starts
to
form
the
tissue
is
sort
of
become
self-contained
and
modular
and
the
question
is:
can
we
characterize
this
formation
of
modules
in
the
embryo,
with
networks
or
network
theory,
and
so
I've
been
working
on
this
for
a
while
with
different?
You
know,
presentations
and
publications
that
are
available
on
this
euler
cycles
for
life.
Talk
is
also
kind
of
included
in
this
because
it
kind
of
deals
with
this
issue
of
emerging
modularity.
C
C
C
We
have
this
experimental
control
network,
which
is
described
in
the
literature
specifically
for
brain
networks,
and
then
you
have
different
types
of
spatial
networks.
So
spatial
networks
turn
out
to
be
important.
Here
we
have
different
types
of
spatial
networks
that
we
can
consider
node
and
edge
construction.
So
this
is
how
you
take
a
bunch
of
nodes
which
are
cells,
and
how
do
you
connect
them
using
what
criterion
but,
more
importantly,
using
what
kind
of
set
of
rules?
C
So
you
know
you
can
use
something
called
preferential
attachment
where
you
attach
to
the
most
highly
connected
cells.
So
you
know
this.
This
sort
of
reinforces
the
idea
of
of
hubs
in
a
network.
Equi-Probable
edges
is
where
you
build
a
random
network,
just
based
on
you
know
any
edge
being
equally
probable.
So
it's
a
random
process
or
a
fitness
model
where
you
look
at
different
connections
and
consider
their
fitness,
you
know.
So
these
are
all
important
issues:
topological
minimization,
so
for
brain
networks.
C
So
you
have
a
number
of
different
ways
to
do
that
and
then
you
know
there's
a
listing
of
some
of
the
main
issues:
the
criticality
in
the
embryo.
We
have
a
couple
of
different
ways
of
looking
at
criticality
wearing
minimization
some
references
there,
network
structure
some
standards,
so
this
is
an
incf
standard
on
standard
representations
of
network
structures
and
then
a
bibliography
which
is
kind
of
what
we
have
here.
But
this
is
a
more
general
bibliography.
C
I
bring
this
to
people's
attention
because
a
this
is
a
kind
of
a
document
that,
if
you're
working
on
an
idea
that
you
can't
seem
to
write
or
before
you
start
writing
a
paper
on
it,
and
you
have
an
idea.
You
have
a
project
that
you
want
to
work
out.
This
is
a
good
way
to
do
this.
You
know
you
start
with
like
some
publication
or
some
presentations.
C
C
This
topic-
and
I
know
yeah,
so
this
isn't
really
a
review,
it's
kind
of
not
really
a
proposal,
it's
kind
of
like
a
way
to
guide
people
through
this
topic,
which
is
actually
quite
difficult
to
understand,
and
we
can
build
on
this.
So
these
are
stubs
that
you
know
you
can
just
say.
Why
is
when
are
wearing
minimization
important?
Why
is
network
structure
important?
C
Maybe
you
know
we
can
consider
these
different
types
of
criticality
in
in
context.
You
know
which
ones
make
more
sense,
which
ones
make
less
sense
to
describe
changes.
So
when
we
have
like
a
neural
network
and
an
embryo
network
when
they
split
when
they
diverge,
is
there
some
sort
of
criticality
that's
driving
this?
Is
there
like
a
neural
avalanche
or
a
jamming
which
is
like
where
you
have
a
bunch
of
particles,
that
kind
of
get
to
a
certain
density
and
then
change
their
phase?
C
Behavior,
there's
a
pro
something
called
propagation
criticality
and
then
symmetry
breaking
all
these
things
may
have
more
or
less
to
do
with
this
phenomena.
We
would
like
to
know
you
know
we
would
like
to
evaluate
those,
but
to
do
those
you
need
to
identify
what
they
are,
and
so
we
have
all
those
things
here,
and
so
I'm
just
pointing
this
out
so
that
people
are
aware
of
this
kind
of
document,
and
I
think
this
is
useful
for
any
idea
if
you
want
to
work
it
out.
C
So
that's
the
link
to
the
document
and
also
that
you
know
if
you're
interested
in
this
topic,
you
know
we
can
talk
more
about
it.
I
think
maybe
this
is
something
that
could
be
of
interest
to
the
neuromatch
crowd,
as
it
deals
with
kind
of
like
emergent,
neural
networks
and
and
things
like
that,
so
the
next
thing
I
want
to
get
to
is
this
okay.
So
I'm
going
to
talk
a
little
bit
about
regulation
and
gene
regulatory
networks
and
before.
C
A
C
C
So
there's
this:
I
wanted
to
get
into
this
topic,
so
this
is
a
book
or
we've
talked
about
gene
regulatory
networks
before
like
the
way
people
have
built
them
in
as
computational
models
and
we've
talked
about
them
in
biology
in
the
context
of
like
genes
being
regulated,
but
there's
actually
this
book
that
is
eric
davidson
wrote
back
in.
I
can't
remember
how
old
this
is.
This
is
not
a
new
book,
but
you
know
the
regulatory
genome,
gene
regulatory
networks
and
development
and
evolution
eric
davidson
is
now
dead
deceased.
C
So
this
is
eric
davidson.
This
is
the
regulatory
genome,
gene
regulatory
networks
in
development
and
evolution.
So
you
can
see
this
big
circuit
here.
This
is
a
gene
regulatory
network
in
the
cell
here-
and
this
is
just
one-
you
know
set
of
genes
that
interact
and
this
so
this
circuit
describes
like
some
process,
that's
being
regulated
for
the
purposes
of
producing
this
phenotype
or
some
phenotype.
C
You
have
a
lot
of
arrows
a
lot
of
lines,
a
lot
of
like
arrows
that
go
up
like
this,
so
this
is
from
2006,
so
his
work-
he
has
been
more-
have
been
working
on
this.
I
think
for
30
years
before
writing
this
book,
and
some
of
the
things
I'd
like
to
point
out
here
is
that
you
have
these
different
chapters
in
the
book.
So
chapter
2
is
inside
the
cis
regulatory
module
control
logic
and
how
regulatory
environment
is
transduced
into
spatial
patterns
of
gene
expression.
C
So
a
cis
regulatory
element
is
where
you
have
the
gene
and
then
you
have
the
regulatory
elements
right
next
to
it,
so
it's
within
a
single
unit
and
cis
and
and
they
talk
about
these
different
principles.
So
it's
very
closely
tied
to
spatial
patterns
of
gene
expression.
So
you
have
these
circuits
that
sort
of
describe
you
know
this
promoter
is
deriving
the
expressionist
gene
and
then
you
know
you
have
this
gene
product.
Sometimes
it
gets
regulated
by
an
enhancer,
so
it's
tuned
its
output.
C
But
all
these
things
are
very
specific
in
space
because
you're
eventually
what's
happening
is
a
phenotype
is
being
built,
and
so
you
know
maybe
cells
in
a
certain
part
of
the
embryo
or,
being
you
know,
have
this
one
expression
pattern
and
in
another
part
of
the
embryo
they
have
a
different
expression.
Pattern
could
also
be
that
you
know
within
the
cell
itself,
there's
spatial
you
know.
C
So
maybe
something
is
expressed
out
at
the
membrane,
but
not
in
the
so
much
in
the
localized
in
the
cytoplasm,
these
sorts
of
things,
and
so
it's
there's
this
downstream
of
specification,
spatial
repression
and
then
the
power
of
the
cis
regulatory
module.
So
this
is
where
you
have
diverse
regulatory
outputs
from
a
simple
input.
C
So
this
is
there's
a
figure
down
here.
Yes,
this
is
it
so
this
is
a
cis
regulatory
control
of
even
skip
the
even
skip,
gene
and
drosophila,
and
the
even
skipped
gene
is
where
you
have
this
pattern:
the
spatially
restricted
pattern
where
you
have
stripes
and
you
have
depending
on
where
the
cell
is
in
this
pattern.
You
have
this
network
where
you
can
have
so
you
have
this
set
of
genes
that
are
sort
of
aligned
on
the
chromosome
and
they're
regulated
in
different
ways.
Some
are
repressors.
Some
are
you
know.
C
Some
are
expressed
highly
so
these
these
stripes
here
the
dark
ones
and
the
repressors
are
the
gray
ones,
and
you
can
see
that
the
activity
is
enhanced
in
these
stripes
and
repressed
in
these
darker
areas,
and
these
are,
I
think
we
talked
about
hox
genes
before
these
are
collinear
on
the
genome,
so
these
are
lined
up
on
the
genome
and
they're.
You
know
expressed
depending
on
their
location
in
the
cell's
location
in
the
embryo,
and
so
this
is
another
example
here
where
you
have
spatial
repression,
regulation
of
this
gene
here.
C
And
so
this
is
something
that
is
a
probabilistic
process,
so
this
doesn't
happen
in
every
embryo
in
the
same
way.
But
it's
you
know,
there's
a
stochastic
component
to
it,
but
it
behaves
in
a
way
that
does
produce
patterns,
and
so
there
are
a
lot
of
ways
you
can
design
these
I
mean
these
are
in
nature,
so
they're
not
designed
they're
actually
evolved,
and
you
can
see
how
these
things
kind
of
co-evolved
over
time.
C
So
there's
this
type
of
so
this
is
a
model
of
cis-regulatory
expression,
drosophila,
the
short-range
repression,
a
through
d,
so
a
through
d
is
where
you
have
short
range
repression
over
space.
You
get
these
spatial
patterns
and
then
e
through
f
are
long
range.
Repression
where
you
get
a
long
range
repression
over
space
and
so.
C
C
So
this
is
a
brief
overview
of
the
genomic
control
apparatus
and
its
causal
role
in
development
and
evolution.
So
there's
this
regulatory
apparatus
encoded
in
dna,
gene
regulatory
functions
and
development,
genetic
genomic
regulatory
sequences
and
the
evolution
of
morphological
features.
So
there
are
a
lot
of
these.
This
goes
back
to
sort
of
this
problem
of
diversity
of
form
and
animal
life.
So,
by
the
end
of
the
50s,
it
was
clear.
C
So
a
large
part
of
the
sensor
lies
in
the
gene
control
circuitry
encoded
in
the
dna.
So
this
is
something
that
again.
This
is
this.
Is
a
graph
of
egg
rna
complexity
and
genome
size,
and
you
can
see
that,
like
in
this
case,
genome
size
as
it
increases
it
sort
of
has
this
relationship
with
rna
complexity,
although
in
the
area
of
10
to
the
sixth
genome
size
in
terms
of
kilobases?
C
There's
a
lot
of
variation
here,
so
there's
some
relationship
between
rna
complexity
in
the
egg
and
genome
size,
although
that's
not
entirely,
it
just
shows
that
you
know
there's
a
lot
of.
You
know
how
much
transcript
is
produced
versus
how
big
is
the
genome,
and
so
the
idea
is
that
there's
a
lot
of
right,
the
more
rna
you
have
in
the
egg,
the
more
regulation
that's
going
on!
So
it's
you
know
it's
that's
what
they're
trying
to
show
there.
C
You
also
have
gene
numbers
and
genome
size,
so
you
see
that
in
general
there
there
are
fewer
genes
than
genome
size,
there's
a
lot
of
genome.
That
is,
you
know
outside
of
coding,
regions
of
the
genes.
So
you
have
a
gene
and
then
you
have
maybe
junk
dna,
or
maybe
you
have
regulatory
elements
and
in
some
species
that's
a
lot
of
stuff
and
in
some
species
it's
less
so
so
there
may
be
some
relationship
between
you
know.
C
C
C
So
this
is
another
example
of
a
regulatory
architecture,
a
modular
cis-regulatory
information
processing
module.
So
this
is
where
they
show
kind
of
the
space
time
domain,
one
space,
time
domain
two
and
then
this
these
other
inputs.
So
you
can
see
that
you
have
these
different
modules
that
are
regulated
by
different
signals.
You
have
different
spatial
repressors
different
signals
from
adjacent
cells
cell
cycle
control,
which
is
something
controlled
by
maybe
some
other
genes
involved
in
cell
cycle,
which
is
a
is
a
cyclical
process
of
different
lineages.
C
You
know
lineage
identities,
which
we've
talked
about
in
our
meetings
with
respect
to
c
elegans,
and
you
have
this.
Essentially
it's
a
mapping
between
these
two
modules.
So
you
have,
like
you
know,
one
module
that
has
one
set
of
inputs.
Another
module
is
a
different
set
of
inputs
or
you
know
their
mirror
inputs,
but
the
point
is
is
that
these
two
modules
will
then
feed
in
to
this
final,
this
final
module
here
and
this
is
an
or
function.
So
this
can
be
like
it
logic
gate,
basically
that
each
of
these
modules
can
be.
C
You
know
you
could
have
one
module
operating
or
another
module
operating
or
both
at
the
same
time
and
there's
an
or
function
here,
which
means
that
it
could
be
either
or
module
feeding
into
this,
and
this,
in
this
case,
they're
all
wired
in
the
same
way
so
they're
basically
redundant,
and
this
is
actually
important
in
evolution,
because
you
know
you
have
like
multiple
modules
that
can
do
the
same
thing.
It
can
step
in
if
one
module
is
say
removed
by
evolution
or
you
know,
there's
a
duplication
of
something,
and
you
know
it.
C
The
circuit
is
sort
of
modified
in
some
way
that
these
redundancies
can
be
very
useful
for
maintaining
the
right
amount
of
rna
output
here.
So
you
know
this
is
also
the
case
in
where
you
have
developmental
mutations
or
ablations.
Where
things
are,
you
know,
maybe
deliberately
shut
off
by
a
experimenter
or
maybe
they're
shut
off
by
some
you
know
developmental
disorder
or
mutation,
and
so
that
in
that
case,
your
the
redundancy
that
you
have
here
is
good.
C
So
there
are
a
lot
of
regulatory
demands
of
of
development,
you're
executing
they
take
the
view
that
it's
the
execution
of
a
genetic
program
that
the
program
is
the
concern
and
the
program
is
maybe
look
something
like
this,
and
so
this
is
the
thing
we
want
to
be
concerned
about
in
understanding
that
wiring
blueprint
and
so
pattern
formation.
We
talk
about
pattern
formation,
a
bit
and
thinking
about
pattern
formation.
C
C
But
people
have
really
struggled
to
try
to
make
these
computationally
salient
and
scaling
them
up
so
yeah.
So
this
is
cis-regulatory
elements
of
some
muscle-specific
genes.
You
can
see
that
they're
high,
a
very
diverse
range
of
elements
that
can
be
used
to
regulate
the
expression
different
of
different
genes
for
different
functions.
So
in
this
case
for
muscle-
and
so
you
can
see
across
species
that
you
have
these
modules,
that
are,
you
know,
kind
of
aligned
on
the
chromosome
that
enable
a
gene
to
be
regulated
in
different
ways.
C
So
you
have
all
these
different
regulators
that
either
can
suppress
things
or
they
can
tune
things
like
a
you
know
an
enhancer
or
they
can
just
you
know,
act
as
like.
A
binary
switch,
they
can
either
turn
genes
on
or
turn
genes
off.
All
those
things
exist
in
these.
This
is
a
very
diverse
type
of
model
for
sort
of
controlling
gene
expression
and
controlling
the
production
of
a
phenotype,
and
so
I
think
that's
all
I
wanted
to
say
about
that.
There's
also
a
review
article
that
I
found
on
this.
C
This
is
eric
davidson
from
2005
and
colleagues.
This
is
computational
representation
of
developmental
gene
regulatory
networks.
So,
as
in
eric
davidson's
later
years,
he
tried
to
implement
this
in
in
celica
and
in
computational
terms,
and
they
actually
did
work
on
this
biotapestry,
which
supports
the
process
of
model
construction
and
also
model
visualization.
C
So
we
have
some
like
and
we've
seen
this
in
the
alef
community
too,
where
people
have
tried
to
make
these
grn
salient
and
a
lot
of
them
are
very
primitive.
You
know
they
either
look
like
this,
which
is
a
hugely
complex
circuit
or
they're,
very
primitive
in
terms
of
what
they
can
do,
but
there's
some
principles
you
can
use
to
build
on.
So
for
one
you
have
developmental
robustness,
which
means
that
you
have
multiple
mechanisms
for
the
same
thing.
You
have
multiple
positive
and
negative
feedback
loops.
C
So
there's
robustness,
you
also
have
other
things
like
redundancies.
So
you
you
know
this
is
where
you
have
multiple
things
that
do
the
same
thing,
but
you
know
it's
like.
If
you
knock
one
out,
it
seems
like
it's
redundant
at
the
time.
You
know.
Maybe
you
have
like
10
genes
producing
the
same
gene
product,
and
you
know
that
may
seem
like
a
waste,
because
a
lot
of
the
product
actually
goes
to
waste.
C
C
You
just
have
you
know
things
just
have
to
they
kind
of
get
generated
and
then,
if
they're
successful,
they're
more
fit
and
then
they
persist
over
time
and
that's
exactly
what
you
see
with
robustness
and
redundancy
is
that
robustness
in
terms
of
function.
Redundancy
is
in
terms
of
sort
of
structure,
and
they
they
give
you
these
options
as
the
organism
moves
through
time,
so
they
they
kind
of
go
through
this.
C
They
say
that
we've
refused
refer
to
the
grn
state
in
a
given
cell
at
a
given
time
as
the
view
from
the
nucleus
in
that
cell,
each
view
from
the
nucleus
describes
the
set
of
genetic
regulatory
interactions
specific
to
a
particular
group
of
identical
cells
at
a
particular
moment
in
time.
So
this
is
an
interesting
way
to
think
of
this.
C
There's
also
the
view
from
the
genome,
which
is
the
summary
of
all
interactions
in
all
cells
of
interest
over
the
entire
period
of
interest,
and
so
vfg
does
not
display
time
or
space,
but
only
has
one
copy
of
every
gene
in
in
any
of
the
vfns
and
summarizes
the
genetic
potentiality
of
the
embryo.
So
there's
this
potentiality
that
we
have
in
the
embryo
to
produce
a
phenotype
and,
in
this
case,
they're
using
these
different
viewpoints
to
characterize
them.
So
you
can
either
have
a
view
from
the
nucleus,
which
is
where
you
have
these
different
interactions.
C
It's
sort
of
more
of
a
cellular
view
and
then
the
view
from
the
genome,
which
is
more
of
a
genomic
view
where
it's
like
looking
at
the
cells
instead
of
the
genes,
and
so
you
put
those
two
together
and
you
get
this.
I
think
we
talked
about
like
sell
regulatory
networks
a
couple
weeks
ago.
We
had
a
paper
on
that
and
then
we
also
had
a
paper
a
long
time
ago
about
grns
in
celica,
where
they
build
models
of
these
genes.
C
I
think
the
step
forward,
though,
is
to
build
something
like
a
gal4
uis
system,
a
computational
version
of
this,
and
if
we
can
do
that,
then
this
is
like
going
to
be
the
breakthrough
in
this
area.
This
is
the
gal4
system.
This
is
a
very
famous
experimental
system.
This
was
developed
in
the
80s
and
90s
actually
were
refined
in
the
80s
and
90s,
where
they
basically
used.
This
is
in
fruit
flies
where
they
use
this
gal4
element
where
you
can
separate
out
like
different
type
different
parts
of
the
gene
expression
process.
C
C
The
benefit
of
a
system
like
this
would
be
in
well
in
biology.
They
use
fluorescent
proteins,
but
you
can,
but
but
the
benefit
of
this
in
a
computational
sense
is
a
divide.
It
partitions
these
two
processes,
it
partitions
the
process
of
activation
and
then
also
the
process
of
what
happens
upstream
in
terms
of
activation.
C
So
it
tells
you
kind
of
like
what's
happening
at
the
enhancer
and
at
the
other
regulatory
elements,
and
then
it
also
tells
you
what
happens
upstream,
what
gets
activated,
and
so,
if
we
divide
it
into
those
two
components
in
a
computational
sense,
we
can
understand
this
process
better.
We
can
build
better
models
of
this
for
different
things,
and
I
know
we
talk
about
genetic
algorithms,
but
this
is
sort
of
a
variation
on
this.
C
Where
you
have
these
regulatory
sequences,
you
know
where
you
have
genes
that
maybe
evolve
through
mutation
or
recombination,
and
then
you
have
these
modules
that
regulate
them
and
tell
you
how
much
of
a
certain
gene
product
you
have
at
a
certain
time
and
then
that
can
lead
to
building,
like
maybe
maybe
simple
phenotypes,
but
to
have
this
type
of
system
in
place.
You
know
you
want
to
be
able
to
like
unpack.
C
What's
going
on
computationally
and
experimentally,
and
this
is
this
is
one
way
to
do
this,
so
I
wanted
to
leave
that
there,
I
think,
that's
a
good
sort
of
pointer
to
the
future.
I
don't
want
to
say
that
this
is
something
that
has
been
solved.
This
is
something
that
needs
to
be
solved.
I
think
it
has
a
lot
of
utility
but
yeah.
I
think
that's
good
and
so
I'll
leave
it
there.