Science 14Feb2020

single-cell state maps and their underlying
clonal relationships.

Clonal dynamics identify early transcriptional
fate boundaries

With LARRY, it is possible to estimate how a
single cell changes over time by sampling a
clone across multiple time points. However,
the accuracy of this approximation depends
critically on the similarity of sister cells at the
earliest time point. We found that pairs of sis-
ters profiled on day 2 localized in the SPRING
graph had correlated gene expression (median
R= 0.846) and that most (70%) fell in the same
or nearest neighbor cluster (Fig. 2, A and B;
fig. S4, a to d; materials and methods, section

5). A minority of cells, however, were more

diverged, with 10% falling outside a four-

cluster radius (compared with 80% for random

cell pairs). We tested and ruled out that similar

sister pairs were technical co-encapsulation

artifacts (fig. S4e). These tests justified approx-

imating single-cell trajectories by clonal tra-

jectories, although with some loss in resolution

of fate boundaries expected because of ~10%

diverged sister pairs.

Beginning with the in vitro data, we re-

corded the clonal fates of each day 2 cell.

Visualizing cells from unilineage clones revealed

well-delineated domains of fate potential (Fig.

2C). Where the progenitors for different fates

overlapped, we observed bipotent or oligo-

potent clones, indicating the location of fate

commitment boundaries (Fig. 2D). The true

number of multipotent clones is likely under-

estimated in our data because some clonal

fates were likely missed due to undersampling

(fig. S5) and cell commitment before division

would result in only one observed fate. Consist-

ent with recent scSeq studies ( 19 ), progenitors

with different fate potentials did not partition

into discrete cell states, but instead formed a

structured continuum. Further, bipotent do-

mains formed extended fate boundaries, indi-

cating that differentiation progression can occur

independently of fate commitment over some

time. Both of these observations differ from the

classical model of hematopoiesis represented

Weinrebet al.,Science 367 , eaaw3381 (2020) 14 February 2020 3of9

Mk, Er, Ma,

Ba, Eos

Lymphoid

Monocytes

Neutrophils

Mk

Er

Ma

Ba

Probability density Eos

K-means clustering

(colors arbitrary)

Day 2 clone (colors arbitrary)

ABJoint probability for sisters D

0 ≥0.8

C

State-fate map of hematopoiesis in vitro

D

Mk,Er,Ma,

Ba,Eos

+Neu

+MoNeu

Mo

+Ly

Bipotent

progenitors

Analysis of sisters cells at day 2

clusters

MPP and GMP

progenitors

Erythroid

progenitors

Basophil

progenitors

Neutrophil

progenitors

Monocyte

progenitors

Dendritic cell

progenitors

B cell

progenitors

T cell

progenitors

Fraction of mature sisters in lineage 0 1

pg pg pg

Rbp1

Esam Ly6a

Hlf

Itga2b

Gata2 Car2

Ly6a

Gp3

Iigp1 Gp3Ly 6 a S100a4

Tie1 Ly6a

IghmFlt3H2afy

Dntt Ms4a3Ctsg

Gene expression enrichment in lineage progenitors: difference of average log expression (in UMIs)

NK cell

progenitors

-log

p-val 10

Flt3

Ighm

Dntt

Gjb3 Fox p 1 Mdm2

E

F

G

Basophil

progenitors

Neutrophil

progenitors

Monocyte

progenitors

Dendritic cell

progenitors

Mast cell

progenitors

Megakaryocyte

progenitors

Eosinophil

progenitors

Lymphoid

progenitors

Gene expression enrichment in lineage progenitors: difference of average log expression (in UMIs)

-log

p-val 10

Podxl

Itga2b

Pbx1Nrgn Lgals9 1810022K09Rik

Ikzf2

Cpa3 Gata2

H2-Q7Srgn Cpa3

Ikzf2

Cyp11a1

Akr1c13

Muc13Srgn

Ccl9Elane

Igfbp4

Rbms1Sirpa

SetTk1

Spp1

Gene expression of functionally primed progenitors in vitro

Gene expression of functionally primed progenitors in transplant hematopoiesis

H

Fraction of mature sisters in lineage 0 1

Ly 8 6

Irf8 Ighm

Mpeg1Olfm1

IghmIl12a

Mef2cFlt3

Bcl11aLmo4

0≥ 10

Gene expression (UMIs)

Cd34 Procr

Gata2 Flt3

Cd34+ cells Procr

Gata2 Flt3

0≥ 10

Gene expression (UMIs)

I

J

Fig. 2. Linking state to fate in early hematopoiesis.(AandB) Sister cells
at day 2 are transcriptionally similar, as seen by example [(A); each color shows
one clone] and by the probability of sister cells occupying the same or
neighboring transcriptional clusters (B). (C) Day 2 cells (colored dots) are
colored by the fate of their mature sisters observed at a later time in vitro.
Outlined regions of the SPRING plot indicate the respective fates. (D) Location of
progenitors (colored dots) with two fates among their sisters at later time
points. (E) Gene expression domains of day 2 cells guide selection of early

progenitors for further analysis. (F) Early progenitors colored by the fraction

of sisters in each fate at days 4 to 6 in culture. (G) Volcano plots identifying

genes enriched among early progenitors for each lineage. Labeled genes are

shown in red. UMIs, unique molecular identifiers. (H) Heterogeneity among

purified LSK cells after 2 days in culture before transplantation into mice. (Iand

J) Detection of early progenitor gene expression associated with future fates

after transplantation, repeating analyses from (E) to (G). In (E), (F), (I), and (J),

points with nonzero value are plotted on top.

RESEARCH | RESEARCH ARTICLE