►
From YouTube: DevoWorm (2023, Meeting #17): GSoC Community, Embryos/ball microscope, Assembly at the edge of life
Description
Google Summer of Code onboarding and welcome. Updates on DevoLearn and embryo segmentation. Community period: getting familiar with OpenWorm, DevoWorm, and Orthogonal Research and Education Lab. Axolotl embryo images from the ball microscope. Assembly theory and the origins of life. Assembly demonstration with Braitenberg Vehicles at the edge of life. Attendees: Sushmanth Reddy Mereddy, Jiahang Li, Himanshu Chougule, Jesse Parent, Susan Crawford-Young, Bradly Alicea, and Richard Gordon.
B
B
A
B
D
D
D
B
D
D
Are
exactly
because
cell
membrane,
segmented
file
should
be
only
segmenting,
the
membrane
of
the
cell,
and
so
here
they
are
extracting
centralization.
We
need
to
extract
the
centroids
in
nucleus,
nucleus
membrane.
Segment
in
nuclear
segment
are
actually
cell
nuclear
segment
that
here
we
need
to
extract
the
centroids
and
the
volume
of
the
cell.
But
here
we
are,
they
are
doing
in
Reverse
method.
I
didn't
understand
that
so
I
was
thinking
in
my
plan
of
action
while
using
second.
D
D
D
D
Data
set
of
cell
tracking
everything
will
be
cleanly
explained,
starting
pre-processing
building
a
model.
Fine
tuning
with
every
different
notebook,
explains
the
different
process
of
these
all
things
and
yeah.
That's
in
one
rep
only
I
will
keep
a
segment
everything
model
I'm,
not
here
against
and
another
tab
on
it.
Also,
in
the
same
rate,
someone
comes
or
reads
the
report
they
easily
can
understand
what
is
going
on
in
the
report.
This
would
be
the
file
structure
and
actually
I
mentioned
that
PG
is
a.
There
is
a
platform.
D
D
C
B
D
So
actually
we
can.
There
is
segmentation
editor
in
this,
with
that
we
can
create
the
ground
Truth
for
every
cell
manually,
and
so,
if
we
create
round
root
for
at
least
10
to
15
cells,
that
would
be
enough
for
training
and
another
five
is
five
other
videos
just
generally
assuming
it
would
be,
but
for
Sam
from
meta.
D
That
is
not
useful
training
at
all,
not
useful,
because
it
was
already
trained
on
11
billion
images,
cells
of
different
things
so,
and
it's
really
not
needed
to
train
it,
but
I
am
thinking
to
train
it
fine
tune
it
for
particularly
to
develop
and
take
out
the
on
index
from
that
and
uploading
face
as
a
app.
We
have
three
models
in
spaces
right
like
that.
Only
I
will
upload
this
instant
segmentation
in
huggling
Phase.
Also,
there
is
actually
support
for
one
next
support:
purpose:
Sam
model.
B
Yeah
so
I
mean
no,
it
looks
good
I
mean
it's
definitely,
you
know
putting
in
different
tools
that
exist
because
everything's
moving
pretty
fast
and
we
always
like
to
improve
upon
what
we're
doing
and
I
like
the
idea
of
using
fiji's
sort
of
like
a
benchmarking
tool.
Eg,
of
course,
is
sort
of
a
version
of
image
J
that
they've
open
source
I
think
they've
specialized
it
for
various
functions
and
things
like
that.
But
it's
yeah.
B
B
If
you
have
some
knowledge
of
what
the
you
know,
what
you're
trying
to
the
questions
you're
trying
to
ask
I,
guess
you
know
we're
we're
kind
of
a
little
bit
further
away
from
that
like
we
want
to
be
able,
to,
you
know,
discover
things
so
yeah
and
then,
of
course,
there
are
all
sorts
of
algorithms
you
can
use
that.
Are
you
know
more
extensive
versus
what
like
PG
and
image
jab
to
offer
so
yeah?
This
is
so.
This
is
your
a
notebook.
This
is
my.
D
Code
actually,
okay,
these
are
things
are
plotted.
Actually,
the
segmentation
masks
and
I
created
this
3D
tool.
Also
to
me
compression
loss.
Is
there
3D
visualization
when
we
convert
a
tip
file
to
gpg,
I'm
just
making
all
this
visualization
tools
to
you
when
someone
visits
it
I
am
adding
some
documentation
to
it.
Also
that
I
am
making
in
a
basis
way
like
starting
to
start.
We
need
to
pre-process
the
data,
a
complete
explanation
of
it
next
segmentation
model,
let's
find
tuning
Etc
if
the
process
will
go
on
like
this.
B
D
Yeah
I
mean
that's
that
is
due
to
fluorescence
effect,
I
think
so
you
see
when
you
see
the
upward
view
of
it.
You
can
see.
This
part
is
like
has
good
Imaging,
but
T
cells
doesn't
have
good
images.
This
is
a
3D
file
right.
When
you
see
from
upwards,
this
is
I
actually
CL
Games
photo.
It
was
seen
an
upper
View,
and
this
is
the
3D
data
set
I
mean
you
can
see
right.
This
kind
of
tables,
yeah.
B
D
I
mean
started,
dividing
from
there
itself,
it
has
the
highest
height
we
can.
This
is
how
we
can
calculate
the
volume
of
cell,
also
because
we
are
having
the
height
area
of
the
Sun
from
upward,
so
we
can
calculate
the
volume
of
each
cell
number
of
frames
into
total
number
of
frames
into
area
of
the
cell,
so
that
would
give
out
the
volume
of
the
cell.
B
D
There
is
no
particular
like
wrong
side,
but
I
was
this
was
mayuk's
idea.
Actually
he
told
to
take
number
of
frames
into
the
area
of
cell
that
could
be
the
volume
of
cell,
because
it's
a
3D
data
set
right.
That
way
we
can
expect
the
volume
of
each
alligator,
so
I
was
trying
to
implement
his
idea
only
and
extract
that.
B
C
A
D
C
D
B
B
C
B
D
B
D
You
have
to
submit
something:
I
actually
got
a
mail.
Actually,
you
need
to
update
in
some
web
app
to
extend
the
time
period
they
have
mentioned
in
the
Key
software.
So
that
is
the
reason
I'm
mentioning
I
want
to
work
as
I'm
a
lot
of
time
on
this
project,
clearly
to
create
the
complete
work
for
everything.
So
I
want
to
take
as
most
how
much
time
I
have
to
to
do
this
project.
Yeah.
B
B
E
E
E
E
E
Things
that
were
mentioned
like
when
sushant
is
doing
six
month
is
doing
on
this,
like
restoring
development
to
hugging
faces,
and
one
of
the
other
things
that
I'll
mention
was
to
keep
develon
as
well
on
the
side,
but
like
whatever
work
we'll
do
this
summer,
we
can
store
like
the
model
weight,
so
we
can
just
import
the
model
to
hugging
face
itself
right.
Another
thing
was
like
since
it's
a
tool
that
can
be
used
by
everyone
and
like
ease
of
uses,
something
which
we
can
keep
in
mind.
E
For
example,
the
current
segmentation
models,
as
well
as
the
model
like
sushma,
that's
going
on
Etc.
So
basically,
there
are
two
ways
they
can
like
work
around.
The
cutting
faces,
like
you
kind
of
just
upload,
the
pth
file,
or
you
can
just
create
a
model
inside
I
mean
faces
like
using
the
way
of
doing
things,
and
both
of
them
is
like
viable
options.
E
E
After
that,
like
one
of
the
things
like
I
was
going
across
was
about
CL
against
itself,
because
it's
something
which
is
new
to
me
and
I
wanted
to
learn
more
about
it.
So
I
watched
some
YouTube
videos
and
one
of
the
things
which
came
for
which
I
came
across
was
someone
Nick,
a
Sydney
Brandon
itself.
They
discovered
a
messenger
RNA
and
they
also
had
solved
genetic
code
and,
after
that,
they
moved
on
to
development
of
the
nervous
system
in
which
they.
B
B
E
B
That
in
years
past
this
part,
the
model
organism
biology,
is
useful,
I
think
but
we'd
always
like
you
know.
We
I
think
we
talked
about
this
last
week.
We
just
like
to
keep
it
up
with
the
latest
stuff.
So
if
you
find
something
that
you
find
particularly
useful,
you
could
add
it
in
there.
If
you
wouldn't
mind.
E
E
B
B
E
B
B
Okay,
so
I
want
to
go
over
some
materials
for
the
community
period,
because
this
is
the
beginning
of
the
community
period.
So
we
have
the
two
people
we've
selected
such
month
and
hamanchu
and
you're.
You
know
this
period
from
now
until
the
end
of
May
is
the
community
period.
So
what
does
that
mean?
And
what
are
we
going
to
do
with
respect
to
our
group?
B
So
the
community
period
is
generally
for
people
to
get
familiar
with
the
community
and
by
Community
I
mean
like
the
the
group
you're
working
with
and
some
of
the
other
resources
that
exist.
So
we
are
the
devil
worm
group
we've
been
around
since
about
2014..
We've
had
collaborators
come
and
go,
and
so
we
have
done
in
in
the
course
of
that
time,
a
large
number
of
collaborations
gsoc
projects
and
the
like.
So
we
have
a
lot
of
things
that
have
been
around
over
time
now.
B
We
also
are
affiliated
with
the
open
worm
Foundation,
so
the
open
worm
Foundation
is
interested
in
the
C
elegans
model
organism.
The
idea
there
is
to
create
a
digital,
C
elegans
with
all
the
different
parts,
so
we're
the
developmental
aspect
of
that,
and
they
have
you
know
the
movement
group.
They
have
the
connectome
group.
You
know
the
physiological
modeling
group
they're
people
who
do
biophysics
and
so
forth.
So
there
are
all
sorts
of
different
groups
within
openworm
that
are
doing
things
publishing
papers
and
the
like.
B
So
there
are
a
couple
resources
I
wanted
to
go
over,
so
the
first
one
is
the
timeline.
So
we
are.
This
is
the
timeline
for
2023
Google
summer
of
code.
We
are
right
about
here,
so
we're
at
the
beginning.
People
were
selected
on
the
fourth
so
which
was
Thursday
and
then
to
if
we
go
out
to
May
28th
from
now
until
May
28th,
that's
the
community
bonding
period.
So
this
is
where
you
can
work
on.
B
B
So
I
know
such
month
has
been
working
in
the
community
here
for
a
while.
So
he
knows
something
a
little
bit
about
it
and
then,
of
course,
hamanchu
has
been
working
a
little
bit.
You
know
he's
kind
of
familiar
with
it,
but
I
would
encourage
you
to
reach
out
I
know
you've
been
exploring
the
onboarding
guide.
B
B
The
reason
I
do
that
is
because
I
want
you
to
understand
what
people
understand,
how
these
things
change
over
time
from
the
proposal
to
the
finished
product.
So
I
want
people
to
do
that,
and
so,
if
you
have
any
questions
about
that,
let
me
know,
but
it's
I
think
the
first
part
is
pretty
straightforward.
B
So
then
we
go
to
August
I.
Think
it's
September,
let's
see
I,
think
actually.
August
28th
is
where
the
last
week,
the
last
week
of
projects
ends
so
the
final
week
gsoc
contributors
submit
their
final
Work,
August,
28th
and
1800
Etc.
So
that's
the
end
of
the
project.
Now
you
can
extend,
you
can
request
an
extension
on
your
project,
which
is
good
because
you
know
sometimes
you
have
things
that
come
up
during
the
summer.
B
B
So
that's
that's
our
schedule.
Then
we
have
the
so
we
have
a
nice
announcement
here,
the
open
source
blog
at
Google.
They
mentioned
that
you
know
there
were
the
acceptances.
Let's
celebrate,
there
were
43
765
applicants
from
160
countries,
7723
proposals
submitted
967
contributors
accepted
from
65
countries.
So
that's
that's
nice
to
see
a
lot
of
worldwide
diversity
there
in
a
lot
of
people
being
part
of
us
this
program
and
then
I
want
to
point
out
the
mentors.
So
jihang
is
a
mentor
for
this
project.
B
He's
here:
Maya
dab
who's,
not
here,
is
also
a
mentor
and
so
you'll
be
going
to
either
myok
or
jihong
for
questions.
Jiahang,
of
course,
is
his
focus,
is
the
graph
neural
networks,
part
and
then
myoc
is
his
part?
Is
the
Devo
Devo
learn
part
he's
sort
of
the
maintainer
for
develer
and
the
reason
we're
doing
it?
That
way
is
because
we
want
to
bring
a
lot
of
stuff
together
this
summer.
You
know
we
want
to
do
the
graph
neural
networks,
part.
B
We
want
to
improve
the
segmentation
part,
and
then
we
want
to
have
like
a
some
sort
of
a
unified
package
where
people
can
come
to
Divo
learn
they
can
do
the
segmentation
in
different
ways.
They
can
do
the
graph
neural
networks,
and
you
know
we
want
to
make
that
usable
for
people,
so
it's
not
like
they
have
to
dig
through
a
lot
of
documentation
to
use
it
and
so
I
hope
by
the
end
of
summer
we
can
get
to
a
point
where
it's
really
looks
like
it's
really.
You
know
coming
along.
B
So
that's
that's
our
goal
for
the
summer,
and
so
that's
great,
and
we
have
so.
We
have
over
2
400
mentors
Jesse
by
the
way
is
also
a
gsoc
mentor
in
different
set
of
projects,
but
congratulations
to
all
the
mentors
I've
prepared
a
blog
post
on
this.
So
this
is
the
Google
summer
of
code
2023.
This
is
a
wild
guy
been
running
for
over
10
years.
It,
you
know,
got
a
lot
of
content
on
it,
but
this
is
just
the
post
describing
our.
B
So
these
are
the
the
projects:
I'm,
mentoring,
I'm,
mentoring
to
students
through
Diva
worm
and
then
I'm
also
mentoring,
people
through
the
orthogonal
research
and
education,
lab
and
Jesse,
of
course,
is
involved
in
that
and
amancio
actually
was
a
gsoc
student
in
the
orthogonal
lab
last
year,
so
he's
actually
been
through
gsoc
already.
So
congratulations
to
everyone.
B
So
this
is
our
community.
Our
community
consists
of
open
worm,
evil
worm
and
then,
to
a
lesser
extent
the
orthogonal
research
lab
so
open
worm
is
this.
This
is
the
website
for
open
worm.
There
are
these
different
projects
like
cybernetic,
which
is
like
a
biophysics
simulator,
so
they
they
can
simulate
the
C
elegans
in
different
environments
in
different
on
different
substrates
and
so
they're.
B
Looking
at
how
the
C
elegant
moves
against
different
substrates
they're,
looking
at
the
biophysics
of
the
organism,
the
muscles
the
soft
tissues
and
the
fluids
that
they
encounter,
so
there
are
also
other
things
like
worm,
Sim
and,
of
course,
Diva
worm.
The
open
worm
browser,
which
is
like
a
three-dimensional
browser
that
you
can
open
up
and
or
a
three-dimensional
viewer
that
you
can
open
up
in
the
browser
and
there's
an
an
initiative
to
bring
this
to
virtual
reality.
B
So
you
know
this
is
a
model
that
kind
of
puts
together
the
morphology
of
the
worm
models,
every
neuron
models,
the
connections.
So
you
can
explore
that
you
can
also
simulate
those
connections
with
something
like
neural
ml.
The
connectome
model-
or
you
know,
Geppetto,
is
a
thing
that
models
individual
cells,
especially
their
electrophysiology,
and
then
there's
other
things
like
c302,
where
they
put
together
put
that
together
in
a
connecto.
B
So
a
lot
of
those
materials
are
in
this
repository,
and
this
repository
is
from
an
open
house
in
2016
that
we
had
where
we
had
everything.
You
know
all
the
different
subgroups
come
together
and
give
presentations,
so
some
of
this
material
was
dated,
but
it
should
give
you
a
good
idea
of
what's
going
on.
So
there
are
all
sorts
of
different
things
going
on
in
the
in
open
worm
from
different
types
of
tools
to
different
types
of
modeling
paradigms
to
different
types
of
approaches.
B
So
I'll
put
this
link
in
the
in
the
slack,
and
you
know
you
can
have
that
as
sort
of
a
way
to
explore
open
worm,
there's
of
course
the
rest
of
the
slack
which,
if
you're
in
the
slack
you're
in
the
diva
worm
Channel
you
go
outside
the
diva
worm
Channel
and
there
are
a
lot
of
different
channels
to
explore.
Some
of
these
are
Project
based.
Some
of
these
are,
you
know,
maybe
organizational
based.
B
There
isn't
a
lot
of
activity
outside
of
the
diva
worm,
Channel
and
the
slack,
but
there's
a
lot
of
stuff.
That's
archived,
so
you'll
be
able
to
learn
something
about
that
Diva
worm
itself,
so
deworm
has
a
website
as
well.
This
is
our
website.
We
have
different
links
here.
We
have
people,
we
have
our
educational
initiatives,
our
media
and
public
lectures,
which
are
different
talks
that
have
been
given
there's.
B
B
B
B
B
So
that's
that's
our
community
overview
so
that
the
community
just
means
people
that
you
might
be
able
to
draw
from
in
terms
of
resources.
So
if
I
go
to
you
know
who
Manchu
wants
to
find
some
resources
on
C
elegans
movement,
you
know
their
resources
in
our
community.
We
can.
We
can
incorporate
they're
data
sets
also,
but
I'm
not
sure
there
was
a
movement
website,
but
I
think
it's
not
no
longer
operative,
so
they're
different.
You
know
there
are
different
people
you
can
connect
with.
B
They
can
provide
different
resources
and
that's
that's
one
of
the
points
of
the
community
period,
the
other
point
of
the
community
period-
and
this
will
be
something
I'll
talk
about
next
time
or
the
time
after
that
is
that
there's
this
whole
issue
creating
open
source
software.
So
at
the
end
of
the
pro
at
the
end
of
your
projects,
you'll
be
creating
software.
It's
open
source
you'll
have
a
license.
You'll
have
some
things
and
then
the
point
here
is
I
want
to
be
able
to
help.
B
C
B
A
A
A
There
can
you
see
that
yeah
yeah,
okay?
Well,
these
are
from
my
my
microscope
and
I
took
several
different,
several
pictures
on
each
View,
sometimes
with
more
lighting,
sometimes
with
less
in
and
out
of
focus
on
various
parts.
So
hopefully
we
can
get
a
good
view
now.
This
is
the
top
at
an
angle
and
I
turned
on
the
light
on
the
microscope,
and
you
can
see
the
top
cells
easier
with
the
light
on,
but
it
obscures
some
of
the
cells.
A
So
then
I
took
images
with
last
Light,
so
you
can
see
where
the
cells
are,
where
the
light
was.
If
you
know
what
I
mean
yeah
and
so
there's
another
view,
some
of
them
are
obscured
by
the
the
surrounding
gel
and
it
might
be
that
I
should
take
the
gel
completely
off
damage
these,
but
I
would
need
a
little
cup
holder,
so
these
are
just
stuck
on
the
slide.
I
just
put
them
on
the
slide
and
took
pictures
of
them.
So.
A
A
Yeah-
and
this
is
where
the
layer
and
development
the
cells
start,
it
starts
Contracting
into
the
bottom.
What
do
you
call
that
I
said
no
anyway,
it
looks
like
it's
in
the
same
position
as
as
these
cells
are
are
dividing,
so
it's
sort
of
interesting
anyway.
This
is
a
better
picture,
but
I
need
to
make
it
larger.
Yeah
I
just
took
pictures.
A
A
Slower
is
more
of
a
problem
with
the
arduinos
and
I
was
running
into
trouble,
getting
them
to
stay
and
focus
some
of
them.
I
guess
it's
just
I'd
have
to
make
sure
the
best
microscopes
were
in
the
slots
and
then
and
then
it
would
work
and
yeah
and
it
changes,
because
the
gel
drives
out
so
I
think
putting
them
in
a
little
cup
holder
would
be
best
so
that
the
focal
range
doesn't
change.
A
A
A
C
A
B
A
A
A
A
A
A
A
A
B
A
A
A
B
A
A
A
A
C
A
A
A
A
A
Baby
I
I,
don't
know
they
get
some
expensive
acquisition
and
they
haven't
got
it
yet
and
stuff.
The
guy
from
can
said
he
wanted
salamander
eggs
to
take
pictures
of
with
the
cans
microscope
and
so
I
sent
him
some
I,
don't
know
if
the
arrive
alive
or
not,
but
I,
don't
think
so.
Okay
yeah
these
are.
These
are
X
novels
from
Kijiji,
they're,
they're,
mine
and
okay.
C
B
I
said
you
know
Manchu
and
such
month
we
should
keep.
You
know
in
communication
on
Slack,
so
you
know
I,
don't
I
don't
expect
a
daily
update.
But
if
you
have
anything
you
want
to
share,
we
have
our
channels.
We
have
our
DMs,
just
just
use,
make
use
of
those,
and
then
you
know
weekly,
we'll
be
having
a
meeting
where
you
should
give
a
short
update
on
what
you've
done
the
last
week
and
then
towards
the
end
of
the
community
period.
We'll
have
a
presentation
where
you
can
talk
about
your
proposal.
B
B
B
A
B
And
now
I'd
like
to
talk
about
a
new
thing,
we're
exploring
within
the
group.
It's
something
called
assembly
Theory,
so
we're
working
on
a
paper
and
some
research
on
the
area
of
early
life
and
one
of
the
issues
in
early
life
is
that
you
have
darwinian
Evolution
that
you
can
trace
back
to
a
certain
point
in
time.
But
before
that
point
things
have
to
assemble
from
non-life,
and
so
the
question
is:
how
does
that
happen
requires
a
new
Theory,
because
we
can't
necessarily
use
darwinian
Evolution
to
describe
the
transition
from
non-life
to
life.
B
On
the
other
hand,
things
don't
just
kind
of
show
up,
they
have
to
assemble
from
basically
biochemistry,
and
so
how
does
that
process
happen?
So
this
is
a
paper
that
I'm
or
it's
actually
an
article
that
points
to
some
papers
that
I'm
going
to
cover
it's
by
Phillip
ball
who's,
a
science
writer
and
the
title
is
a
new
idea
for
how
to
assemble
life.
This.
E
B
We
want
to
understand
complex
construction
such
as
ourselves,
assembly
Theory
says
we
must
account
for
the
entire
history
of
our
such
entities
came
to
be,
and
so
you
can
see,
there's
this
network
of
Lego
bricks
in
this
image.
That
sort
of
explains
what
they're
trying
to
get
at
they're
trying
to
get
at
this
idea
that
things
are
assembled
at
least
a
minimal
living
system,
and
then
the
minimal
living
system
can
evolve
from
there.
So
they
have
this
assembly
of
this
little
Lego
person.
B
B
But
in
terms
of
plausibility,
only
a
sub,
a
certain
subset
of
objects
are
possible,
and
indeed
those
are
the
ones
that
emerge
to
become
life.
So
not
every
combination
is
equally
plausible
life
on
other
worlds.
If
it
exists,
might
be
so
alien
as
to
be
unrecognizable.
There's
no
guarantee
that
alien
biology
would
use
the
same
chemistry
results
on
Earth
with
familiar
building
blocks
such
as
DNA
and
proteins.
B
Scientists
may
even
spot
the
signatures
of
such
a
waveforms
without
knowing
they're
the
work
of
biology.
This
problem
is
far
from
hypothetical
in
April
the
European
space
agency's
juice,
spacecraft,
blasted
off
from
French
Guiana
on
a
course
to
Jupiter
and
moons.
One
of
these
moons
Europa
has
a
deep
Briny
ocean
beneath
the
Frozen
crust
and
is
the
most
among
among
the
most
promising
places
in
the
solar
system
to
look
for
alien
life,
and
so
basically,
the
search
for
life
suffer
assume
a
fatal
flaw,
which
is
there's
no
definite
signature
of
life.
B
So
you
can
search
for
the
signatures
of
life
we
know
of
on
Earth.
We
can
go
to
deep
sea
vents
and
look
at
some
of
the
things
that
are
going
on
there.
So
we
can
look
at
a
number
of
different
things
on
this
planet
in
terms
of
the
life
that
exists
on
this
planet.
But
if
we
go
to
another
planet,
those
processes
might
drive
be
driven
by
different
chemistries,
and
so
we
don't
necessarily
know
what.
B
B
So
this
was
led
by
Lee,
Cronin
and
Sarah
Walker.
They
propose
a
very
general
way
to
identify
molecules
made
by
living
systems,
even
those
using
unfamiliar
chemistries.
Their
method
simply
assumes
that
AOE
and
life
forms
will
produce
molecules
with
a
chemical
complexity
similar
to
that
of
life
on
Earth,
and
so
in
last
week's
meeting
we
talked
about
the
Miller
urray
experiment,
where
they
applied
an
electric,
an
electrical
field
to
a
bag
of
chemicals,
a
chemistry
that
had
some
potential
and
they
were
able
to
generate
a
lot
of
different
macromolecules
that
were
life-like.
B
And
so
this
experiment
has
been
done
and
again
and
again
so
we
know
that
early
life
had
a
lot
of
electrical
potential,
a
lot
of
heat
energy,
a
lot
of
lightning,
and
so
we
know
that
that
was
abundant
in
that
environment,
and
these
chemistries
were
also
abundant
in
different
places
and
so
put
that
together,
a
new
good
life
and
probably
starting
multiple
times,
and
it's
not
clear
necessarily
how
that
coal
less
than
the
tree
of
life.
But
this
is
the
problem
we're
going
to
face
in
another
on
another.
D
B
Say
that
they
would
have
a
chemical
complexity,
there
would
be
some
event
that
gives
birth
to
life,
probably
multiple
such
events,
and
then
that
emerges
into
some
sort
of
tree
of
life
for
that
planet
called
assembly
Theory.
The
idea
underpinning
the
pair
of
strategies
even
grander
aims,
so
they
have
a
series
of
Publications.
B
They
have
like
three
Publications.
They
link
to
here
the
first
one
being
this
one:
exploring
and
mapping
chemical
space
with
molecular
assembly
trees.
This
isn't
science
advances.
This
gives
an
idea
of
this
sort
of
combinatorial
space
and
what
that
looks
like
there's
another
paper
here
that
is
exploring
the
sequence
base
of
unknown
ligamers
and
polymers.
So
this
is
again
exploring
that
combinatorial
space
and
then
the
third
one
is
formalizing
the
pathways
to
life
using
assembly
spaces.
B
And
so
this
Theory
attempts
to
explain
why
apparently
unlikely
things
such
as
you
and
me
even
exist
at
all,
and
it
seeks
that
explanation,
not
in
the
usual
manner
of
physics,
in
Timeless
physical
laws,
but
in
a
process
if
it
imbues
objects
with
histories
and
memories.
But
what
came
before
them?
It
even
seeks
to
answer
a
question
that
perplexed
science,
scientists
and
philosophers
for
Melania.
What
is
life
anyway?
B
B
So
the
principle
here
is
that
these
plausible
paths,
through
the
space,
are
driven
by
constraints
but
they're
also
driven
by
opportunities,
and
so
we
don't
really
know
how
these
Pathways
emerge.
We
just
know
that
some
are
going
to
be
more
plausible
and
others,
and
that's
based
on
the
factors
unique
to
that
setting.
So
on
Earth
there
may
have
been
a
certain
set
of
conditions
that
led
to
the
emergence
of
Life
on
another
planet.
You
might
have
another
set
of
conditions,
they.
B
To
Earth
they
can
be
different
planetary
history,
the
chemistry
of
the
life
form,
those
are
all
different,
and
so
that's
going
to
not
only
yield
a
different
combinatorial
space,
but
a
different
set
of
plausible
pathways.
So
life
on
Earth,
for
example,
may
not
be
very
plausible
in
a
place
like
Europa,
but
it
still
may
have
life.
B
B
B
In
this
set
of
Pathways
arises
this
assembly
index,
so
how
do
you
assess
these
Pathways
and
their
success
or
their
failure?
And
so
one
way
to
do
this
is
to
use
this
assembly
index.
This
is
for
various
substances
by
repeatedly
measuring
their
molecular
structures.
So
you
look
at
the
molecular
structure.
B
So
we
have
a
non-biological
realm.
We
have
coal,
quartz
and
granite,
and
so
these
are
coal
comes,
of
course,
from
organic
materials.
Quartz
and
granite
do
not,
but
they
should
have.
You
know
you
would
think
what
is
the
complexity?
What
how
do
you
I
mean?
You
could
have
a
common
framework
to
compare
them?
How
do
they
compare
so
coal
actually
has
so
low
here
is
10
to
12.
So
there's
like
a
range
of
values
here,
it's
like
this
is
the
low
value.
This
is
the
high
value.
B
This
is
the
average
they
just
give
the
numbers
of
all
on
the
high,
so
coal
is
a
low
of
10,
but
I
have
12.
quartz
and
so
remember.
Coal
is
derived
from
organic
materials.
Quartz,
on
the
other
hand,
has
a
low
of
11
and
a
high
of
12.,
which
means
that
you
have
this,
that
the
quartz
may
be
a
little
bit
more
complex.
B
Granite,
however,
has
a
low
of
10
and
a
high
of
15
with
an
average
of
12..
So
that
means
that
Granite
is
maybe
even
more
complex
than
Kohler
quartz,
despite
being
a
Rock
versus
like
something
that
derived
from
organic
materials
or
another
type
of
rock.
So
you
know
clearly
there's
some.
You
know
variation
here,
but
these
are
all
non-biological
biology,
so
they
don't
really
ex.
You
know
they're,
not
living
systems.
B
Now
you
can
look
at
living
systems
as
well,
and
so
here
we
have
yeast
urine
seawater,
E
coli
and
beer
and
of
course,
beer
is
derived
from
yeast.
So
there's
this
aspect,
but
again
like
coal,
while
being
derived
from
a
living
system
or
biological
system
organic
system.
So
we
can
see
here
that
all
these
biological
systems
are
clearly
higher
in
complexity
than
the
non-biological
ones.
So,
if
you
compare
it
with
rock
with
stone,
something
like
yeast
has
a
low
of
10
and
a
high
of
24
with
an
average
of
14..
B
So
you
see
that
the
low
and
the
average
assembly
index
aren't
really
that
much
different
from
non-biological
systems.
But
you
do
get
these
high
levels
of
complexity.
That
and
I
don't
know
how
they
exactly
measure
this,
if
they're
measuring
behavioral,
States
or
different
conformational
states
of
proteins
or
what.
But
we
can
see
that
there's
a
higher
high-end
measurement
here,
the
same
thing
with
urine,
which
is
urea
and
water,
and
things
like
that,
where
you
go
from
nine
to
27
with
an
average
of
14..
B
It
also
carries
waste
when
it's
suspended
in
water
and
so
forth,
seawater,
which
of
course,
has
a
lot
of
microorganisms
in
it.
It's
you
know
not
organic
and
of
itself,
but
it
has
a
lot
of
Harbors
a
lot
of
Life
and
DNA
and
so
forth.
Seawater
has
a
low
of
9
high
of
28
and
an
average
of
13..
So
it's
somewhat
like
you're
in
some
like
yeast
E
coli
is
of
course
organism.
B
You
know
you're
sampling
and
we're
a
single
cell
organism
with
DNA
proteins
and
the
like,
it's
probably
a
little
more
concentrated
than
what
you'd
find
in
your
inner
seawater
urine,
also
carrying
some,
maybe
some
microbial
life,
but
E
coli
has
a
low
of
15,
a
high
of
31
and
average
of
20..
So
clearly
this
is,
you
know
we
have
a
living
organism
versus
Rock.
E
B
Beer,
which
is
derived
from
yeast,
has
a
low
of
10,
a
high
of
34,
an
average
of
16..
So
these
high
measurements
are
really
interesting
because
for
living
systems,
they're
very,
very
high
compared
to
non-living
systems,
and
so
you
know
the
question
arises.
You
know
what
what
is
contributing
to
this,
especially
when
we
see
that
the
low
values
are
all
sort
of
similar.
What
does
it
arise
from,
and
so
this
is
really
I
think
showing
that
the
assembly
index
is
measuring
something
about
biological
complexity
and
livingness.
B
Okay,
now
I'm
going
to
talk
about
one
of
the
papers,
in
particular
that
we're
linked
to
from
that
quanta
article
and
I'm,
going
to
compare
it
to
some
of
the
work
we've
been
doing
with
the
breitenberg
vehicle
toy
models,
so
this
paper
is
exploring
and
mapping
chemical
space
with
molecular
assembly
trees,
and
so
they
use
a
tree
of
steps
to
show
the
path
through
this
possibility
space.
So
we
have
an
example
here
in
figure,
one
where
you
have
representations
of
an
assembly
pathway.
So
you
see
that
step.
B
One
is
where
you
have
these
molecules
that
are
dissociated
step.
Two
is
where
they
come
together
in
a
bond
step.
Three
is
where
they
have
another
Bond
come
together
and
then
step
four
is
where
it
forms
an
adenine.
So
you
get
these
different
steps
where
things
are
assembled,
and
then
this
leads
to
some
sort
of
thing,
that's
biochemistry
from
a
simple
chemistry.
So
you
have
this.
B
You
know
you
have
this
toolbox
of
things
they
get
put
together
in
stages,
and
you
end
up
in
this
situation,
where
you
have
a
biomolecule
after
a
number
of
assembly
steps.
So
there
are
these
Pathways
that
come
through.
You
know
that
are
more
plausible
than
others,
given
the
nature
chemical
bonds,
the
the
different
electrical
charges
and
so
forth.
So
so
the
colors
show
these
building
blocks.
C
B
The
Shield
building
blocks,
you
have
a
toolbox
building
blocks,
you
put
them
together
and
things
get
assembled
a
certain
pathway,
so
that's
showing
basically
what
they're
talking
about
when
they
say
assembly,
Theory,
you're,
assembling
things
from
smaller
parts.
So
this
is
a
molecular
assembly
tree
where
you
have
this
path
through
a
possibility
space.
There
are
a
number
of
possibilities
for
assembling
things
and
you
get.
The
path
still
doesn't
really
show
the
tree
explicitly.
B
But
if
we
go
down
to
this
figure
figure
three,
that's
a
little
bit
clearer.
You
have
these
building
blocks.
You
have
these
paths
to
different
things:
ADP
ATP
and
AD,
plus
different
bases
of
a
RNA
molecule,
actually
a
DNA
molecule.
Then
you
have
RNA
and
thymine
uracil.
So
all
right.
So
this
is
the
four
bases
of
an
RNA
molecule.
You
have
a
uracil
gets
concerned
thymine,
and
these
are
the
four
bases
of
the
DNA
molecule.
So
you
can
see
how
that
works
and
again
these
are
all
necessary
for
a
living
system.
B
B
And
so,
if
we
turn
to
auditory
model,
I've
worked
out
an
example
using
breitenberg
Vehicles
how
this
might
work
with
their
toy
model.
So
we
have
these
assembly
Pathways.
We
have
different
rules
of
assembly.
This
is
we
had
in
the
assembly
tree
and
we
can
see
how
this
works.
So
you
have
a
vehicle
body,
a
single
panel
of
a
vehicle
body.
It's
you
start
with
that.
You
start
with
a
number
of
those.
You
add
another
part.
So
there's
a
rule
that
says
you
can
duplicate
the
single
panel,
so
you
have
two
panels.
B
Those
get
put
together,
as
you
see
at
the
end,
and
then
you
have
these
wheels,
which
are
the
four
things
on
the
outside
those
get
put
together
and
those
two
steps,
duplication
and
addition
you're
duplicating
this
part,
you're
adding
these
parts-
and
you
get
this.
This
is
a
viable
vehicle
because
it's
by
some
by
laterally,
symmetrical
as
it
doesn't
have
too
much
space
where
there's
no
weight
bearing
aspect
to
it
and
it
can
Propel
itself.
B
So
in
this
case,
what
I'm,
showing
you
here
is
the
viable,
come
for
these
assembly
rules
and
these
assembly
steps.
But
let's
suppose
that
that's
not
the
case.
Let's
suppose
that
we
have
an
asymmetry,
so
we
have
a
panel
and
we
have
a
rule
of
implication.
So
you
have
two
panels
that
kind
of
align
top
to
bottom
or
I,
guess
anterior
posterior,
and
then
you
have
not
four
wheels
but
three
wheels.
B
So
this
is
this
circle
with
a
slash
thread,
but
the
thing
is:
is
that
if
it's
close
enough
to
being
viable,
you
might
actually
have
the
opportunity
to
rescue
something
so
so
this
is
all
non-living
s.
Things
are
added
together
using
your
rules
and
we
end
up
with
this
non-viable
vehicle.
But
it's
non-viable.
It's
not
completely
non-viable,
it's
somewhat
non-viable
and
so
it's
possible
to
rescue
that
function
using
a
flagellar
rotor,
which
we've
added
here
in
step.
C
B
This
is
another
addition
Rule
and
now
you
have
a
viable
vehicle,
and
so,
according
to
our
Criterion,
it's
now
functional
and
it's
not
chemistry,
but
it's
this.
This
vehicle
toy
model
example
where
you
get
this
transition
from
non-living
to
living,
where
in
this
case
it's
non-functional
to
functional,
and
so
this
is
just
an
example
moving
from
sort
of
the
hardcore
biochemistry
to
something
that
maybe
straddles
living.
You
know
simple
living
organisms,
slash
behaving
organisms,
so
just
wanted
to
show
you
that.