There are over 37 trillion of them in the human body,

100 times more than the number of stars in the Milky Way.

So each of the cells in our tissues

fulfills a different type of role,

but together they like make this beautiful symphony

that lets the tissue maintain itself

and then lets our organs do their functions

and eventually our entire body.

Over the past 300 years,

we've learned what cells are made of,

how they function and divide into new cells.

But there's still a lot we don't know.

How many different cell types are there in the human body?

How do different cell types work together?

And how do changes in cells cause diseases?

About 2,000 researchers from over 70 different countries

are trying to answer these questions

by building a Human Cell Atlas,

a complete map of all the cell types in our body.

What we envision is that the Human Cell Atlas

will be a foundational reference for biomedical research

in many, many areas.

And it tells us something about ourselves,

about what our bodies are made of,

and also going to help us and many others

over time develop new medicines for patients.

So we like to think about the final atlas fondly

as a Google Map of the human body.

You could say, well, I want to understand the cells

in the nose, or the cells inside your mouth,

or the cells of the skin,

or the cells of this particular region in the brain.

And you could drill into that region.

First you could look at coarse resolution.

You would know how the tissue is ordered,

and then you could go in finer and finer resolutions

all the way to the level of individual cells.

Aviv Regev is the co-chair of Human Cell Atlas,

a 10-year endeavor to discover new cell types

as well as map them in detail.

The other way that we like to think about the atlas

is what we call the periodic table of the cells.

If you think about the periodic table of the elements,

it's not just a description of the elements.

It's also a theory of the elements.

You know, Mendeleev was able to predict

that elements would exist

before they were actually empirically found.

And one of the things that we hope will happen

with this atlas is that we will learn

how to better predict cells and their behaviors.

In 1664, around 200 years before Mendeleev

and his periodic table, English scientist Robert Hooke

discovered the existence of cells

when he put a piece of cork under a microscope.

Ever since then, microscopy has played an important role

in studying cell structure and function.

By looking at cells under a microscope

and studying their reactions

with chemical stains, which make them visible,

scientists identified about 300 cell types

in the human body.

But cells which look similar under a microscope

can sometimes turn out to be chemically different.

And so our knowledge has been limited until now.

Something happened a few years ago,

which was a major technological advance

that allows us to look at the molecular content

of the individual cells

through their RNA molecules in particular.

And we call this single-cell genomics.

In the past, we take say a piece of tissue

that would have many different kinds of cells in it.

And we would put it basically

through the lab equivalent of a blender.

And so if you think of every cell

as a different piece of fruit,

there's blueberries and strawberries

and raspberries and kiwis and so on,

then what you get as a result of that is very similar

to a fruit smoothie.

It's a blend of all of the molecular contents

of all of those cells

and you get to measure the average.

That is not a great way by which to discover

what are the individual cells.

What single-cell genomics allows us to do

is look at every individual cell

in the molecules within it separately.

And so this gives us basically the equivalent

of a fruit salad.

Now you can see each individual piece of strawberry

and blueberry and kiwi and raspberry.

Where the workflow starts is with acquiring the tissue.

And so in that case there are biopsies

from deceased transplant donor tissue.

The biopsy tissue is broken down

into single whole cells.

These individual cells are loaded

onto a microfluidic droplet robot,

which carries out the chemical reactions

needed to prepare them for sequencing.

We're measuring which genes are switched on in each cell.

To understand why single-cell

sequencing is important,

we need to understand a little bit about how cells work.

Our genome, which is made of DNA,

is the instruction manual for the cells in our body.

Within this genome are thousands of different genes,

which each code a different protein.

These proteins are made using a chemical called RNA.

By using single-cell sequencing,

if you can identify which genes

each individual cell is using,

you can tell what sort of cell it is.

But there's one problem.

We have about 25,000 different genes in our genome.

And with single-cell genomics,

we can measure several thousand per cell.

And so what the technology is telling us

is in each single cell which specific subset

of 2,000 or 3,000 genes is switched on in that single cell

out of the 25,000 possible.

For every single cell, you have several thousand genes

expressed and you can have hundreds of thousands of cells.

So the data matrix is hundreds of thousands

multiplied by thousands of data points.

So we really are talking about sizes of data

that are getting close to astronomical.

With the help of machine learning

and artificial intelligence,

these huge amounts of data can be processed and analyzed,

eventually leading to the discovery of new cell types.

The Human Cell Atlas community has been able to reveal

dozens, maybe now up to even a hundred,

different new cell types

across different tissues of the body.

So for example, several years ago

together with my colleagues,

we did a study of the airways in the lungs

and we discovered a new cell type,

which is very rare and that nobody knew existed.

It was literally not there until we discovered it.

Of course it was in our lungs,

but it was not in our knowledge.

And that cell type that we call the ionocyte

ended up being especially exciting for us.

Because it uses, it expresses very highly

the gene known as CFTR, or the cystic fibrosis gene.

And until we discovered this very rare cell type,

scientists actually assumed that this gene

is used by other cells in the lung and the airways.

And it turns out that it is not used by those cells.

It is used by these super rare cells

that we didn't know existed.

Cystic fibrosis is a fatal hereditary disease

that affects the lungs and digestive system.

The discovery of this new cell type

could help diagnose and treat the disease.

Our hope is that this would allow us

to also understand disease better.

Because what happens in disease is that cells

all of a sudden misbehave.

They don't do the things that they're supposed to do.

And what we want to develop

are the right kinds of medicines of course

that would bring them back to their native state.

And we hope that the atlas would be a foundational

reference and resource like the human genome has been

in order to help scientists both understand

basic biology and understand disease.

Individual cell types identified

using single-cell sequencing

can then be located within tissue samples

to make a detailed 3D map.

So far over 39 million cells have been analyzed,

covering specific organs, such as the brain, skin and lungs.

But to make a complete cell atlas covering every tissue,

organ and system in the human body,

billions more cells need to be analyzed.

It is a crazy quest and it's ridiculously ambitious,

but we have these little milestones along the way.

And each time we discover something new,

there's an incredible excitement and thrill.

And that keeps us going as well.