►
From YouTube: DevoWorm (2021, Meeting 16): Patterns in Networks/Seashells, Digital Bacillaria, Embryos/Organoids
Description
Review of Seashell Pattern Recognition, Embryo Networks, Connectomes, and Euler Paths for Life, Digital Bacillaria review, Fish Ladder model of Abiogenesis, Optic Cups in Vertebrate Embryos and Organoids, Plant Non-neuronal Cognition, and Spatial Differentiation in Organoids. Attendees: R Tharun Gowda, Jesse Parent, Richard Gordon, Susan Crawford-Young, Krishna Katyal, Vrutik Rabadia, Bradly Alicea, Mainak Deb, Vash Yadi, Shruti Raj Vansh Singh, and Akshay Nair.
A
A
A
C
C
C
A
C
C
C
C
And
so
that's
a
good
place
to
start,
but
all
we
need
is
multiple
images
and
a
small
rotation
of
the
snail
in
each
step,
and
that
will
give
us
full
coverage
and
then
we
can
just
take
you,
take
the
closest
lines
and
you
stitch.
You
only
take
the
vertical
a
row
of
pixels
or
or
a
strip
of
them,
and
you
stitch
them
together:
okay,
great
okay
and
since
you're
measuring.
A
B
C
Yeah
but
yes,
he's
very
interested
in
the
problem,
but
we're
way
ahead
of
him
yeah
and
I
discussed
with
him
the
the
problem
of
the
inverse
problem
of
figuring
out
what
the
goals
are.
Yeah
he's
interested
in
that
problem
and
he's
thinking
about
it
in
the
neural
network
context.
In
other
words,
how
do
you
go?
How
do
you
look
at
the
results
of
a
neural
network
and
figure
out
rules
from
it
so
he's
working
on
that
problem,
but
he
does.
He
said
it
explicitly.
He
has
not
solved.
C
C
A
C
C
A
C
A
It
has
the
elongated
cells
on
the
front
and
the
enlarged
cells
on
the
back,
and
that
seems
to
be
how
it
moves
through
the
embryo.
But
I
think
that
it's
changing
cells,
where
it
doesn't
tell
me
enough,
like
the
paper,
so
though
we're
just
treating
it
as
though
it's
self-propelled,
and
it
goes
through.
The
embryo.
D
C
B
C
C
C
Well,
if
you
do
that,
then
there's
a
probability
going
backwards
and
there's
a
step
there's
a
little
bit
of
a
hill.
You
have
to
go
over
to
go
backwards,
so
so
each
step
been
a
stable
state
and
the
idea
here
is
that
to
go
through.
What's
the
probability
of
getting
all
the
way
to
the
top
for
a
molecule,
and
it
turns
out,
if
you
do
around
10
to
the
40th
experiments
you
can,
you
can
get
up
20
steps.
C
C
C
G
B
Yeah
yeah,
okay,
okay,
go
ahead
with
your
agenda;
okay,
yeah
so
well!
Well!
Welcome
to
the
meeting
I
see
some
of
you
filtered
in
so
this
is.
I
just
wanted
to
open
the
floor.
If
anyone
I
have
wanted
to
present
anything
today,
any
people.
I
know
a
couple-
people
in
slack
contacted
me
about
maybe
presenting
something.
H
A
H
H
A
H
B
H
B
G
I
B
B
B
I
don't
know
might
not
clarify
there.
Surety
says
of
course,
we'll
not
be
able
to
present.
Today
was
a
very
hectic
week.
Glad
you
showed
interest
actually
says.
Yes,
it
is
not
perfect.
I
have
to
make
some
changes
so
all
right,
so
that's
great.
Actually
could
you
stop
sharing
your
screen?
It's
frozen.
I
think.
G
B
B
B
H
E
B
All
right,
then
krishna
says
you
will
lose
data
points.
I
guess
with
the
segmentation
aspect,
but
yeah
then
dick
says
want
to
track
each
cell
cell
divisions
change
of
neighbors
differentiation
waves
versus
time,
so
we
you
know
we
want
to
have
different.
We
want
to
be
able,
at
the
end,
to
have
different
types
of
data
that
we
can
collect
or
analyze
the
the
projections
that
we're
making.
So
keep
that
in
mind
when
you're
developing
the
technique
area.
E
E
B
B
If
not
I'll
start
in
with
my
stuff
that
I'm
going
to
talk
about
here,
okay
and
then
actually
it
says
here
could
be
done,
we'll
add
it
to
goals
and
then
dick
says,
where
possible
also
label
the
cell
as
to
type
yeah.
So
these
are
all
things,
of
course,
that
we
should,
if
you're
doing,
project
3.3
to
keep
in
mind.
B
All
right,
so
the
first
thing
I
want
to
talk
about
today
was
I
just
wanted
to
put
in
a
plug
again
for
neuromatch
academy.
We
talked
about
that
last
week.
They've
been
the
course
content
here
on
the
this
is
for
the
summer
school
for
computational,
neural
neuroscience,
and
so
this
has
all
of
the
materials
in
here.
B
They
they
have
the
syllabus
here.
So
it's
a
three-week
program.
You
learn
a
lot
of
different
things
in
that
three
weeks.
It's
a
action-packed
time
and
they've
developed
all
the
different
materials,
so
you
can
actually
look
at
them
here,
tutorials
projects
and
a
welcome
video
and
they
go
through.
You
know
what
the
different
things
that
they
you'll
encounter
during
this
summer
school.
So
there's
a
primer
on
python
for
nma,
which
is
the
academy
name,
and
so
you
get
a
refresher
on
python
and
they
have
those
those
are
pre-work
or
pre-summer
school.
B
B
Then
week
two,
you
go
into
different
types
of
statistical
models,
so
you
have
bayesian
statistics,
linear
models,
decision,
making
optimal
control,
reinforcement,
learning
then
in
week
3
they
talk
about
neurons
and
networks
and
causality
and
then
finally
deep
learning
in
the
last
two
days,
and
so
I
know
jesse
did
this
last
year.
He
can
tell
you
all
about
it.
If
you
want
to
know
more,
that's
you
know
it's
like
one
of
these
intensive
workshops
where
you
spend
like
six
days
during
the
week.
B
So
that's
and
then,
of
course,
as
I
said,
I
think
last
week
they
have
a
not
only
a
computational
neuroscience
course,
but
this
year
they
have
a
deep
learning
course,
which
is
just
devoted
to
deep
learning.
They
have
a
lot
of
speakers
here
that
are
high
profile
and
then
they
don't
have
the
course
materials
up
for
that
yet,
but
they
basically
have
an
outline
of
a
of
a
syllabus
here,
so
you're
going
to
talk
about
a
lot
of
different
types
of
models,
optimization
regularization,
multi-layer
perceptrons.
B
B
It's
pretty
interesting,
so
I
wanted
to
also
then
move
on
from
that
and
talk
about
this
networks
conference.
So
this
networks,
2021,
is
it's.
This
is
the
this
is
the
conference
where
we
got
a
mayonnaise
or
abstract
accepted
on
embryo
networks
and
connectomes,
and
I
think
I
showed
this
a
while
back
this.
This
abstract-
and
this
is
going
to
have
to
be
like
a
12
to
15
minute
presentation.
B
B
One
is
the
neuro
network
neuroscience
and
the
other
is
something
called
topo
nets,
and
this
is
the
submission
for
topo
nets,
which
is
this
euler
cycles
for
life,
and
I
presented
on
this
a
while
back-
and
I
did
a
lot
of
you
know-
work
on
this.
This
is
just
me
on
it
and
I
kind
of
worked
it
out
into
a
two-page
extended
abstract.
C
Can
you
make
a
comment?
One
of
the
things
in
stephen
wolfram's
presentation
was
that
he
he's
going
after
the
problem
of
continuing
representation
of
embryos
and
he's
doing
this
by
taking
each
note
in
a
graph
such
as
you
have,
and
replacing
each
node
by
a
small
graph
which
represents
things
like
the
reproduction
of
itself.
H
B
C
B
Yeah,
okay,
so
yeah!
This
is
well.
This
isn't
exactly
that,
but
the
way
that
this
works
is
that
you
have
these
morphologies
and
you
have
these
networks
and
you
don't
worry
about
the
nodes.
So
much
as
you
do
the
edges
and
then
you
draw
like
what
they
call
an
euler
path
through
each
edge
and
you
try
to
evaluate
the
topology
as
a
as
a
whole
topology
or
if
it
breaks
into
different,
like
modules.
B
So
if
you
can't
complete
an
euler
path,
that's
sort
of
like
the
foundation
of
a
module
and
these
networks
grow
over
time.
So
you
can
add
to
the
networks
and
see
how
they
you
can
evaluate
them
at
each
step
to
see
how
they
grow
and
differentiate
and
so
and
yeah.
I
talk
a
little
bit
it's
two
pages,
so
you
can't
get
very
much
into
that
space,
but
I've.
I've
also
shown
a
tree
where
it's
broken
down
into
a
tree
where
their
growth
rules
applied,
and
you
expand
this
this
shape.
B
B
You
know
I
give
a
talk
on
and
I
can
sort
of
flush
that
I
you
know
some
of
that
out
a
little
bit
more.
So
this
is,
you
know,
an
iterative
process,
so
yeah.
I
think
that's
interesting.
There's
a
lot
to
do.
I
think
with
networks-
and
you
know
other
types
of
mathematic
mathematical
models
here,
so
yeah,
probably
yeah.
That's
what.
D
C
B
C
B
So
that's
that's
the
that
submission,
and
so
let's
see
that
works
so
yeah
we'll
talk
more
about
that
in
the
coming
weeks.
Then
I
wanted
to
point
this
tool
out.
I
don't
know
if
people
have
seen
this,
but
this
is
something
called
archivist,
and
so
this
is
something
where,
if
you're
interested
in
a
certain
topic,
you
can
search
for
a
field
in
this
search
terms,
and
then
it
gives
you
information
about
all
the
recent
preprints
on
different
topics.
So
this
is
like
you
can
search
by
field.
B
I
search
for
developmental
biology,
but
there
are
many
fields
you
can
search.
You
can
have
all
all
repositories,
you
can
search,
bioarchive,
meta
archive
or
maybe
even
archive,
and
then
you
know
over
time
and
then
I
don't
know
if
these
are
the
most
well.
I
guess
these
are
the
most
downloaded
preprints,
so
I
don't
know
if
you
can
set
it
towards
something
else,
but
I
think
that's
in
any
case
that
gives
you
this
list
of
popul.
B
You
know
things
that
are
popular
among
people
looking
at
them
and
downloading
them
and
oftentimes.
These
are
like
different
versions
of
things
that
are
already
published,
so
it
can
lead
you
to
things
that
are
trending
in
the
literature,
so
you
have
all
these
different
things
that
have
been
downloaded
say
in
the
last
30
days,
and
so
I
think
this
one
we
covered
last
week
or
one
of
these
we
covered
last
week.
B
B
Resolved
single
cell
out
atlas
of
human
gastrulation,
that's
a
pretty,
looks
like
a
pretty
important
paper
selling
in
3d
by
incremental
deep
learning.
That's
something
that
is
probably
interesting
to
a
lot
of
people
here.
So
you
know
you
can
learn
a
lot.
This
is
a
good
way
to,
I
think,
search
the
literature
and
find
out
what's
trending,
I
mean
there
are
other
tools
you
can
use
as
well,
but
this
is
something
I
wanted
to
point
point
out
to
people.
B
B
B
It
would
be
really
helpful
if
both
of
you
had
deeper
insights
in
what
is
required,
and
so
dick
put
a
reference
in
here.
Epithelius
bubble
wraps
a
new
method
for
measuring
cell
shape
and
intrasolar
adhesion
and
embryonic
and
other
epithelia,
that's
from
1982.
So
that's
another
paper
that
might
be
useful
for
that
project.
3.3
and
I
think
that's
it
so
that's
side
conversation
so
yeah.
So
these
are,
you
know,
there's
a
lot
of
interesting
stuff
in
here.
B
So
let's
see
I
don't
want
to
get
into
this.
Yet
I
think
I'll
talk
about
the
submissions.
Next.
Actually,
no
I'll
talk
about
this.
I
just
wanted
to
point
out
that
the
digital
basilaria
repository
has
a
couple
of
people
who
have
contributed
this
week.
So
thank
you
to
josh
and
thank
you
to
throne
for
submitting
things
into
the
repository
and
getting
pull
requests
accepted.
B
I
know
josh
has
been
contributing
a
lot
to
the
repository.
Surety
made
a
a
pull
request
that
was
accepted
a
while
back,
and
the
roon
has
also
done
this
as
well
so
and
other
people
who
are
submitting
applications
for
gsoc
have
me.
You
know.
In
the
last
month,
they've
made
some
good
pull
requests
that
have
been
accepted.
B
A
B
Two,
so
this
is
the
digital
basilaria
project,
so
I
mean
only
one
person
is
going
to
get
selected
and
we
may
or
may
not
even
get
that
slot.
So
I
just
want
to
prepare
you
for
that,
but
I
think
there's
a
lot
of
stuff.
We
can
work
with
here.
If
people
want
to
continue
this
work,
we're
very
open
to
this,
and
I'm
not
really
sure
how
this
is
going
to
be
organized.
C
Area
yeah.
Well
I
I
put
the
the
old
book
that
you
wrote
in
2005
on
basil
area
in
the
in
the
chat
now.
One
thing
we
included
in
the
bibliography
are
many
videos,
digital
videos,
online
on
vessel
area
and
I'm
sure,
there's
been
quite
a
change
in
those.
Since
then,
some
have
gone
out
and
probably
a
lot
more
come
in
if
we
can
get
a
volunteer
to
just.
C
B
B
C
B
Yeah
yeah,
so
that,
like
there
are
a
lot
of
we
actually
for
the
the
paper
we
put
out
in
2019
2020
this.
That
paper
had
a
lot
of
that.
We
drew
from
some
of
the
movies
that
are
online
and
you
know
the
there
are
things
on
youtube
where
people
will
take
microscopy
videos
of
bacillaria,
which
is
a
type
of
diatom
and
they'll,
look
at
it
over
time
and
they're
varying
quality.
So
that's
a
an
issue.
C
C
B
B
B
C
Data,
let
me
make
a
a
suggestion:
thomas
harbinger
is
having
trouble
with
keeping
the
cells
investilary
in
focus.
Okay.
Now,
if
you
look
at
a
movie,
if
an
individual
cell
goes
in
and
out
of
focus,
perhaps
one
could
identify
the
in-focus
image
and
then
use
it
to
backtrack
to
the
places
where
it's
out
of
focus
and
substitute
the
in-focus
image,
yeah,
okay
and
then
you
might
produce
you.
Might
you
words
using
time
you
might,
since
the
shape
of
an
individual
cell?
C
C
B
Yeah
yeah,
so
there
are
a
lot
yeah.
There
are
a
lot
of
things
we
can.
You
know
tricks,
we
can
do
on
that
so
yeah
and
then,
as
for
the
paper
that
I
talked
about
last
week,
I'm
sort
of
still
in
the
process
of
getting
that
together.
So
there
needs
a
lot.
More
content
needs
to
be
added
and
a
lot
more
organization
needs
to
be
done.
I
was
going
to
talk
about
it
more
today,
but
I
didn't
get
a
chance
to
get
it.
B
B
B
If
they
have
a
deadline,
let
me
know-
or
I
guess
maybe
jesse's-
usually
the
one
to
do
a
lot
of
editing
on
it,
so
maybe
send
it
to
him
or
okay.
So
you
have.
Let
me
go
back
to
the
chat
here.
So
dick
put
that
basil
area
citation
in
the
chat
yeah.
She
says
I'm
not
able
to
contribute
for
two
weeks
due
to
spring
semester
exams.
B
B
A
B
Retired
no
exams
yeah.
So
so
let
me
go
back
to
the
submissions
this
coming
week.
There's
this
evolution
2021,
I'm
not
sure
I'm
going
to
submit
this
euler
paths
for
life
at
evolution,
because
I've
already
submitted
it
to
the
networks
conference
which
might
be
a
better
fit.
So
I
think
I'm
going
to
gray
that
out
and
then,
as
for
I
don't
know
if,
if
krishna
still
was
wanting
to
do
this
kill
the
winners
idea,
but
the
deadline
for
this
is
coming
up
this
week.
B
So
if
you're
gonna
do
it,
you
know
pull
the
trigger.
But
if
you
don't
do
it,
then
you
can
probably
submit
it
somewhere
else.
We'll
talk
about
that
the
evil
and
paper.
That's
long
suffering-
and
I
know
I
haven't
touched
it
in
a
while-
and
we
haven't
talked
about
it,
but
that's
still
probably
going
to
happen.
Maybe
we'll
wait
a
little
bit
on
g
sock
again
since
we're
doing
some
revisions
or
we're
doing
some
development
on
the
platform
and
the
program
itself.
So
you
know
I
don't.
B
This
will
be
something
maybe
over
the
summer.
That
will
hopefully
happen.
We
have
a
version,
that's
shorter,
but
I
wanted
to
make
it
a
longer
version.
So
then,
this
non-neuronal
cognition
paper-
I
just
updated
you
on
that.
This
is
there's,
obviously
an
extension
on
this
that
I
negotiated.
So
this
is
going
to
be
probably
later
on
in
the
spring
or
summer,
when
this
is
done.
B
This
international
c
elegans
conference,
abstract
we
haven't
heard
back
on
that
this
is
the
let's
see.
Well,
we
have
these
different
things
that
are
like
accepted
and
out.
That's
in
the
green.
These
orange
things
are
things
that
have
been
submitted
and
some
have
been
accepted,
some
not
so.
This
is
the
one
growth
growth
form
and
the
theory
of
deep
learning
that
was
submitted
to
net
neurosatellites.
B
B
That's
we're
still
waiting
on
that,
although
they're
supposed
to
let
us
know
any
day
whether
that's
been
accepted
or
rejected,
and
we
have
things
like
the
boring
billion.
This
idea
for
a
book
contribution
the
kindle
book
idea,
mathematics
and
diva
warm
ideas.
There
are
a
lot
of
things
hanging
out
there.
I
just
want
to
remind
people
those
things
so
that,
if
you're
interested
in
pursuing
it,
you
know
it's
kind
of
front
of
mind.
B
Then
this
is
the
euler
cycles
for
life.
This
was
submitted
to
a
satellite
called
topo
nets
and
again
this
was
just
submitted.
So
I
don't
know
if
this
has
been
accepted
or
rejected.
Yet
we're
waiting
on
that
and
then
there
are
some
other
things
coming
up.
This
testa
williams,
symbiosis,
molecular
level
simulations
of
diatom
motion
jerkiness.
B
Those
are
all
things
that
are
sort
of
on
the
list,
and
so
then,
finally,
you
know
if
some
of
the
machine
learning
oriented
people
are
thinking
of
nurips,
the
submissions
are
in
may
and
they
have
other
workshops
and
satellites
that
have
later
deadlines.
So
that's
pretty
much
all
for
the
submissions
document
I
did
talk
about
digital
basil
area
talked
about.
B
So
that's
good,
but
there
are
a
lot
of
things
on
here
where
you
know
maybe
they've
fallen
through
the
cracks
or
maybe
they're
things
that
we
need
to
update
with
respect
to
what's
going
on
in
the
group,
so
you
know
there.
I
don't
know
about
this
fish
ladder.
This
is,
I
think,
maybe
maybe
in
progress
or
I
guess,
finished
now
or
yeah,
we'll
submit.
I
B
B
Finally,
I
wanted
to
get
to
the
papers
because
I
mean
you
know.
Usually
we
have
a
lot
shorter
period
of
time
for
this,
but
I
know
some
of
you
have
to
leave
at
the
top
of
the
hour.
So
I
got
a
number
of
papers,
it
sort
of
backlogged.
Here
I
got
a
couple
from
susan
this
week
she
sent
a
couple
of
papers
along
and
then
I
found
a
couple
papers.
So,
let's
see
where
would
we
start
here?
B
G
B
Or
something
like
that,
so
this
is
that
starts
out.
I
propose
to
consider
the
question:
can
machines
think
so?
This
is
something
you
don't
necessarily
think
you'd
start
out
with
in
plants,
but
this
should
begin
with
definitions
of
the
meaning
of
the
terms
machine
and
think
so
that's
you
know.
Obviously,
with
this
proposal
alan
turing
moves
into
the
question
whether
intelligence
is
bound
to
neurons.
B
It
is
remarkable
that
he
begins
a
seminal
work
on
that.
What
today
is
so
right
away
and
inappropriately
called
artificial
intelligence,
and
so
they
kind
of
get
into
this
idea
of
non-neural
cognition
they
ask.
Do
you
even
need
neurons
for
intelligence,
and
so
you
know
some
people
might
think
of
intelligence
as
sort
of
what
they
call
anticipatory
systems,
and
we
know
that
there's
been
work
by
robert
rosen
on
anticipatory
systems
with
respect
to
biology.
B
So
you
know
there
are
already
these
connections
with
biology
and,
with
you
know,
regulation
in
the
biological
world
that
are
sort
of
at
least
consistent
with
the
idea
of
intelligence.
So,
if
you
think
of
intelligence,
very
broadly,
you
know,
do
you
really
need
neurons
or
is?
Is
it
something
else?
Is
it
actually
that
neurons
are
a
means
to
an
end.
B
Yeah
yeah
this
is
so
this
is
kind
of
goes
through
this
question
and
and
kind
of
going
through
whether
plants
can
feel
pain,
for
example,
or
whether
plants
are
endowed
with
consciousness,
and
you
know
this
is
of
course
you
know.
These
are
things
that
are
hard
to
sort
of
prove
and
hard
to
get
a
scientific
caster
on,
but
these
are
things
that
you
know.
If
you
think
of
intelligence.
Broadly,
you
can
at
least
begin
to
answer,
and
so.
B
Yeah
and
that's
another,
a
number
of
slime
molds,
I
think
as
well.
They
have
a
lot
of
a
lot
of
the
collective
behavior.
Literature
really
examines
like
collectives
of.
Maybe
you
know
single
cell
organisms
or
insects
or
some
you
know
and
then
says:
well,
there's
this
intelligence
that
supersedes
the
individual.
It's
like
these
collective
behaviors
that
are
quite
complex,
and
so
that's
another
way
to
think
of
it
too.
B
Yeah,
so
so
this
this
paper
goes
on
about
about
developing
a
turing
test
for
plants,
so
there's
an
art,
their
arguments
against
the
plant
neurobiology,
and
so
this
is
the
the
con
part
of
this,
where
they
kind
of
say.
Well,
you
know,
there's
not
really
a
neurobiology
of
any
sort,
so
plant
share
with
animal
cells,
as
with
all
other
cells,
the
ability
to
generate
membrane
voltage
and
can
unders
under
certain
circumstances,
even
generate
action
potentials
and
there's
also
a
vascular
system
that
transports
water
but
also
signals.
B
But
then
this
gives
you
this
network
throughout
the
organism.
However,
without
a
phenomena
is
whether
a
phenomenon
is
a
signal
or
a
bi
phenomena,
it
depends
on
its
meaning.
So
then
they
get
into
this
notion
of
whether
that
network
actually
transmits
information
like
a
brain
might,
and
so
this
is
this
kind
of
goes
through
a
little
bit
of
this,
and
so
this
is
this
paper
kind
of
goes
through
a
lot
of
these
arguments.
Plants
can
plants
can
anticipate.
B
I
don't
know
if
it's
necessarily
it's
like
three
pages,
so
it's
not
really
delving
super
deeply
into
this
topic,
but
they
kind
of
go
through
a
lot
of
the
different
things
like
they
go
through
the
mcculloch
pitts
model
as
a
metaphor
of
neural
learning,
so
you
know
to
think
about
it
in
terms
of
those
type
of
models.
This
is
actually
a
lot
of
the
stuff.
That's
going
to
come
up
in
the
bacillary
non-neuronal
cognition
paper.
B
B
There
is
a
stimulus
response
in
plants,
but
is
that
a
sort
of
a
consciousness
or
not-
and
you
know
what
does
that
mean?
Does
that
mean
that
the
is
there
a
turing
test
that
you
can
develop?
That
says
whether
plants
are
self-aware
or
not
self-aware,
but
maybe
that's
not
even
the
best
way
to
answer
that
question.
Maybe
it's
just
a
matter
of
seeing
whether
they
have
this
ability
to
process
information.
B
B
B
The
basic
structure
of
the
eye,
which
is
crucial
for
visual
function,
is
established
during
the
embryonic
process
of
optic
cup
morphogenesis.
So
this
is
your
eye.
It's
sort
of
the
cup
that,
where
the
eye
is
going
to
form
molecular
pathways
of
specification
and
patterning
are
integrated
with
spatially,
distinct
cell
and
tissue
shape
changes
to
generate
the
eye.
So
you
have
the
specification
in
patterning:
that's
spatially
distinct,
so
you
have
different
things
happening
in
different
parts
of
the
cup
that
I
have
to
happen
in
that
order
in
that
spatial
order.
B
In
order
for
the
thing
to
function,
so
they
have
discrete
domains
and
structural
features,
retina
and
retinal
pigment,
epithelium
and
wrap
the
lens
and
the
optic
cup
fissure
occupies
the
ventral
surface
of
the
eye
and
optic
stalk.
So
again,
these
things
have
to
be
in
the
right
place
and
they
have
to
occur
at
the
right
time
in
development
for
the
the
eye
to
be
built.
C
B
B
But
you
have
this,
so
this
process
has
to
be
sort
of
in
place
so
interest
in
the
underlying
cell
biology
of
eye.
Morphogenesis
has
led
to
a
growing
body
of
work,
combining
molecular
genetics,
so
this
and
imaging
to
quantify
cellular
processes.
So
you
can
look
at
things
like
cell
adhesion
and
what
they
call
ectomyosin
activity,
which
are
these
sort
of.
B
These
studies
reveal
that
intrinsic
machinery
and
spatially
temporally,
specific
extrinsic
inputs
collaborate
to
control
dynamics
of
cell
movements
and
morphologies
during
this
process.
So
there's
this
intrinsic
machinery-
and
there
are
these
cues
from
the
outside-
that
work
together
to
control
this
process.
Here
we
consider
recent
advances
in
our
understanding
of
imorphogenesis,
with
a
focus
on
the
mechanics
of
eye
formation
throughout
vertebrate
systems,
so
not
not
octopus,
but
your
vertebrate
systems,
including
insights
and
potential
opportunities
using
organoids,
which
may
provide
a
tractable
system.
B
So
they
sort
of
get
on
the
organoid
bandwagon
here,
and
so
this
paper
goes
through
sort
of
the
steps
in
development.
So
there's
a
vagination,
which
is
something
that
happens
during
the
formation
of
several
tissues
and
organ
organs,
which
is
this
process
of
folding
and
there's
also
invagination
and
elongation.
B
B
So
yeah
because
well
they
go
through
a
lot
of
different
things.
Here
they
go
through
optic.
Fissure
morphogenesis
is
a
multi-step
process
of
formation
enclosure,
so
they
show
this
kind
of
operation
formation
and
then
fissure
closure.
So
you
have
these
things
and
they
all
have
to
be
timed
just
right.
B
B
I
don't
know
if
they
talk
about
well,
we
talked
about
it
just
now,
with
the
you
know
actually
trying
to
see,
but
they
don't
get
into
that
part,
but
that's
also
a
thing
that's
going
to
affect
it.
As
well,
so
they
kind
of
go
through
this,
and
I
think
the
there
are
no
more
figures
in
the
paper,
but
I
think
the
so.
The
process
of
vimorphogenesis
is
a
good
model
for
looking
at
biological
mechanics.
B
B
If
we
can
improve
our
measurement
or
imaging
and
visualization,
we
can
understand.
Maybe
somehow
these
mechanistic
processes
are
conserved
across
vertebrates
or
even
organoid
systems
which
they're
using
in
this
example.
In
a
couple
of
these
examples
and
again
the
organoid
systems
are,
you
know,
from
like
say
a
human
cell
or
a
mouse
cell
outline,
but
they
actually
develop
on
their
own,
given
cues
in
a
in
a
in
sort
of
a
fixed
environment,
and
you
can
see
parallels
between
the
organoids
which
are
outside
of
an
organismal
context
and
then
like
a
developing
embryo.
B
Oh,
I
don't
want
to
do
that.
One
yeah,
I
think
I'll
do
this.
So
I
know
some
of
you
have
to
leave
now,
so
I'm
going
to
talk
about
one
more
paper
and
then
we'll
wrap
up.
So
this
is
called
lineage
reporting
reveals
dynamics
of
cerebral,
organoid,
regionalization,
and
so
this
is
a.
This
is
also
a
newer
paper.
This
was
from
last
year.
B
I
don't
know
if
it's
in
a
journal
right
now,
but
so
the
the
abstract
here
is
diverse
regions,
develop
within
cerebral
organoids
generated
from
human
induced
pluripotent
stem
cells,
which
are
these
cells
that
they
can
put
transcription
factors
in
and
turn
into
a
stem
cell
from
a
differentiated
cell
like
a
skin
cell.
So
they
can
generate
these
cerebral
organoids.
B
In
this
way
they
need
a
stem
cell
and
then
they
can
grow
these
organoids
and
then
they
you
know
they
can
do
all
the
things
you
can
do
with
organoids,
so
they
they
have
these
organoids
and
they
find
that
diverse
regions
develop
within
them.
So
you
have
these
stem
cells
that
this
population
that's
expanding,
and
you
have
this
these
sort
of
this
differentiation
within
the
organoid
itself.
B
Inducible
crispr
cast
scarring,
which
is
another
molecular
technique,
and
single
cell
transcriptomics,
where
they
actually
look
at
like
gene
expression
in
single
cells,
analog
lineage
relationships
during
cerebral,
organoid
development.
So
now
they're
going
to
analyze.
You
know
what
lineages
these
cells
are,
and
so,
if
you
know
anything
about
c
elegans,
you
know
that
c
elegans
has
a
lineage
tree.
That's
well
worked
out,
but
in
in
organoids
in
in
these
kind
of
regulative
environments,
where
there
are
a
lot
of
regulatory
signals
shaping
the
differentiation
of
cells,
you
have
these.
B
B
In
terms
of
where
they're
coming
from
so
they're,
coming
from
a
stem
cell
and
they're
differentiating-
and
the
question
is,
is
you
know
what
is
the
pattern
of
differentiation?
Do
they
start
off
with
like
two
distinct
types
and
then
differentiate
into
more
specialized
types
or
do
they?
You
know
you
know
what
is
the
pattern
there?
So
that's
what
they
want
to
know.
B
They
observe
increased
fate
restriction
over
time,
which
means
that
there's
a
restriction
of
fate
in
a
cell
over
time,
so
it
becomes
more
and
more
specialized
and
find
that
these
induced
stem
cell
clones
used
to
initiate
organoids,
tend
to
accumulate
in
distinct
brain
regions,
so
they're
actually
able
to
find
that
the
the
initiating
clones
the
clones
from
the
initiating
cells
sort
of
accumulate
in
distinct
parts
of
the
of
the
brain.
So
it's
I
don't
know
what
that
means,
but
it's
an
interesting
finding.
B
We
use
lineage
coupled
spatial
transcriptomics,
which
is
again
this
looking
at
gene
expression
in
certain
parts
of
the
organoid
to
resolve
lineage
locations,
as
well
as
confirm
clonal
enrichment
and
distinctly
patterned
brain
regions.
Then
they
use
this
4d
light
sheet
microscopy
to
track
the
cell
nuclei.
So
this
is
what
a
lot
of
you
are
doing.
You're
tracking
the
nuclei
of
the
cell-
or
you
know,
there's
usually
like
a
fluorescent
signal
in
the
nucleus
and
so
tracking.
The
nucleus
allows
you
to
track
different
cells.
It
serves
as
a
proxy
for
cell
position,
but
it's
not.
B
No.
It
doesn't
tell
you
anything
about
the
full
cell.
It
just
tells
you
where
the
cell
is
and
identify.
You
know
individual
cells.
They
were
able
to
do
this
in
the
organoids,
and
in
doing
so
they
can
link
brain
region
clone
enrichment
to
positions
in
the
directed
room,
which
is
a
layer
of
the
developing.
B
You
have
this
more
advanced
structure
and
you
have
so
you
have
this
process
going
on
early
and
then
you
have
this
later
process
where
things
are
more
specialized
and
they're
limited
in
terms
of
their
migration.
So
our
data
shed
light
sheds
light
on
how
lineages
are
established
during
brain
organoid.
Regionalization
and
our
techniques
can
be
adapted
to
any
sort
of
induced
stem
cell
culture
system
to
dissect
lineage
alterations
during
some
sort
of
you
know:
disrupt
environmental
disruption
called
perturbation
or
on
some
sort
of
patient-specific
model
of
disease.
B
So
these
organoids
a
lot
to
study
diseases,
and
you
can
use
this
sort
of
method
to
look
at
how
diseases
unfold
in
these
organoids
and
and
and
it's
a
an
analog
for
development.
Although
imperfect,
so
they
go,
it's
a
very
technical
paper
and
I
know
that
working
through
that
abstract
was
very
hard.
Maybe,
but
they
have
some
really
nice
images
here.
They
have
a
umap
decomposition
of
the
cells
here
by.
B
I
think
this
is
by
gene
expression,
assay
and
they
show
the
identity
of
different
neurons
in
different
regions
of
the
data
space,
and
so
you
can
see
that
there's
a
differentiation
of
different
types
of
neurons
here
according
to
this,
and
they
just
show
the
different
types
of
cells
that
they
have
so
yeah.
There
are
a
lot
of.
B
And
then
this
is
like
days
of
time
over
over
time.
So
you
start
out
with
the
ectoderm,
then
the
epithelium,
and
then
you
get
into
these
fully
formed
organoids
and
then
at
30
days,
post
plating
these
into
this
culture.
Then
you
get
this,
you
do
the
rna-seq
and
you
look
at
this
and
you
get
this
sort
of
map.
B
I
mean
there
are
a
lot
of
pat.
There
are
a
lot
of
nice
figures
in
here,
so
I
don't
know
maybe
they're
more
figures
than
like
you
know
understanding,
but
that
happens
with
a
lot
of
high
profile
papers.
I
imagine
this
is
in
one
of
the
major
journals
by
now
but
yeah,
so
I
want
to
just
want
to
show
how
they
do
some
of
these
things.
You
know
they're
looking
at
like
different
like
clustering
techniques
and
then
trying
to
infer
like
a
like
a
lineage
tree
from
this.
B
So
you
have
a
lot
of
different,
so
these
cells
get
mixed
in
in
you
know:
it's
not
like
c
elegans,
where
they
have
a
very
distinct
location
that
they're
going
to
end
up
in.
They
there's
a
lot
of
mixing
and
there's
a
lot
of
local
signaling
that
influences
what
these
cells
become,
but
you
nonetheless
trace
the
lineage
and
get
different
types
of
lineage
trees.
Reconstructed
from
this,
it's
a
much
harder
process.
B
B
B
B
Is
the
middle
of
the
organoid
in
the
middle
of
the
organ
there's
a
different
environment
from
the
outside?
Okay?
So
that's
good
octopus
eye
references.
This
is
these
are
on
the
octopus
eye.
We
mentioned
earlier
a
lot
of
them
actually.
So,
if
you're
interested
in
that,
what
is
the
title
of
the
paper?
Which
paper
was
that.