SUPPLEMENTARY METHODS Patient material selection criteria

SUPPLEMENTARY METHODS

Patient material selection criteria

The search term was “metastatic adenocarcinoma” and the years included were 2000 through

2011 (N =3823). To be included for further analysis, the patients had to be deceased, have

detailed clinical data on primary cancer, axillary metastasis as well as distant metastasis

available, and enough paraffin embedded material to enable exome sequencing, gene

expression and immunohistochemical stains from each site. Core and fine needle biopsies

were not eligible for inclusion. In total twenty patients met the criteria. Formalin fixed

paraffin embedded (FFPE) tissue sections were retrieved from all lesions. From the majority

of primary cancers and metastases, multiple tumor areas of different topography were isolated

(>5 mm distance from each other) resulting in 104 samples. Five metastatic samples (two

samples of bone relapse in patient 6, one region of local recurrence sample in patient 13 and

two samples of bone relapse in patient 12) failed during exome sequencing due to insufficient

DNA, resulting in a total of 99 samples.

Tissue microarray (TMA) and IHC staining

FFPE sections were conditioned in CC1 solution (Ventana Medical Systems, Tucson, AZ,

USA) for 36 min (Ki67) to 64 min (PR) and incubated with mouse monoclonal antibodies for

Ki67 (clone Mib-1) (Dako A/S, Glostrup, Denmark) and rabbit monoclonal primary

antibodies (Ventana) for ER (clone SP1), PR (clone 1E2), and HER2 (clone 4B5) at 35 °C

(HER2) or 37 °C (others) for 16 min (Ki67) to 44 min (ER) according to the manufacturer’s

instructions, and finally counterstained with hematoxylin. Two independent pathologists

(NFM and GS) at Karolinska Institutet performed scoring of ER, PR, HER2 and Ki67 and the

consensus values were used to determine IHC-based surrogate subtype for each cancer

sample. The assessments of ER, PR, HER2 and Ki67 IHC were combined into surrogate

subtypes using definitions recommended by expert recommendations (1-3). For a laboratory

specific threshold for Ki67 in Tissue Microarray (TMA) specimens, we incorporated digital

image analysis of a previously published cohort (n=130) of consecutive cancer specimens

collected at the Department of Pathology, Uppsala University Hospital, Uppsala, Sweden

from January 1 1987 through December 31 1989 (4-7). Surrogate subtype classification

based on IHC is illustrated below:

Luminal A-like: ER ≥ 1 % and PR ≥ 20 %. HER2 “negative” and Ki67 < 4.1 %.

Luminal B-like: ER ≥ 1 % or PR ≥ 1 % and HER2 “negative” and Ki67 ≥ 4.1 %, or

ER ≥ 1 % or PR ≥ 1 % and HER2 “positive” Any Ki67or ER ≥ 1 % and PR < 20 % and

HER2 “negative”. Any Ki67.

HER2-enriched-like: ER < 1 % and PR < 1 %. HER2 “positive”. Any Ki67.

Basal-like: ER < 1 % and PR < 1 %. HER2 “negative”. Any Ki67.

PAM50 molecular subtyping after subgroup-specific gene-centering

PAM50 molecular subtyping (8) of each tumour sample was performed after subgroup-

specific gene-centering (9). The population- based Stockholm cohort with primary breast

cancer patients (10) (GEO:GSE1456) was used as training cohort. The subgroup of patients

with breast cancer relapse within the first five years was used to mimic the tumour

progression cohort. All molecular subtype analysis was done in R/Bioconductor.

The PAM50 centroids and Entrez Gene IDs in the pam50 data object in the package genefu

was used. The hgu113a.db and hgu133b.db annotation packages were used for the Stockholm

data and 49/50 PAM50 genes had mapped probesets on the Affymetrix HG-U133A and HG-

U133B arrays. For probesets that were present on both arrays, the average value was used.

For probesets that were mapped to the same Entrez Gene ID, the one with highest

interquartile range was selected. For each PAM50 gene, the subgroup-specific percentile of

the global median in the training cohort was identified. The value 50 (i.e. the median) was

imputed for the one gene (KRT17) where gene-expression data was missing.

In the tumour progression cohort, all PAM50 genes have mapped probesets on the Affymetrix

Human Transcriptome Array (HTA) 2.0 platform (GEO:GPL17586) as given by the

manufacturer’s annotations. Again, for probesets that were mapped to the same Entrez Gene

ID, the one with highest interquartile range was selected. The baseline expression of each

gene was assigned at the subgroup-specific percentile of the breast samples in the tumour

progression cohort (median aggregated by patient). Thereafter expression data for each

sample was gene-centered by subtracting the baseline expression.

For each sample in the tumour progression cohort data, the Spearman’s rank correlation

between the sample after subgroup-specific gene-centering and each of the five PAM50

subtype centroids was calculated and the class of the most highly correlated centroid was

assigned to the sample. Finally, a stringent criterion of nearest centroid correlation coefficient,

larger than 0.25, was applied to assign a final subtype classification.

Main assumptions in Dollo parsimony

We used a variant of parsimony-based phylogenetic reconstruction method named Dollo

parsimony to reconstruct phylogenetic tree for each patient. We used Rdollop() from R

package Rphylip, which uses the implementation “dollop” given in PHYLIP version 3.696.

Following are the main assumptions in Dollo parsimony:

1. We know the state of each ancestral site (in germline) to be 0.

2. The sites (mutations) evolve independently.

3. Each lineage in the phylogenetic tree evolve independently of each other.

4. Probability of acquiring a mutation, i.e., changing from state 0 to 1 is small.

5. Probability of a losing a mutation (a deletion), i.e., changing from state 1 to 0 is also small,

but still far greater than the probability of acquiring a mutation.

Validation of phylogenetic trees

We validated the phylogenetic trees produced by Dollo parsimony using two approaches.

First, we performed phylogenetic reconstruction by an orthogonal method “LICHeE” v1.0

(60), We used the following parameters: -minVAFPresent 0.05 -minClusterSize 10 -

maxClusterDist 0.25 -maxVAFAbsent 0.

Second, we validated that the phylogenetic trees are not affected by variable coverage and/or

different tumor purity between samples. We adapted a modified approach from Yates et al

(16) to identify and remove mutations whose presence or absence in any sample from a

patient is indeterminate due to either read coverage or lower tumor purity, i.e., they can be

missed by chance. Then, phylogenetic trees were reconstructed using Dollo parsimony after

removing all indeterminate mutations. Supplementary Figure 11 contains a side-by-side

comparison of the trees in each patient.

To identify indeterminate mutations, we computed the upper 95% confidence interval (CI) of

VAF for each absent mutation in each sample according to the binomial distribution. If the

upper 95% CI exceeded a threshold VAFthr, the mutation was marked as indeterminate.

VAFthr is defined for each mutation as the maximum observed VAF for that mutation in other

samples from the same patient multiplied by the ratio between tumor purity in the sample

having the maximum VAF and tumor purity of the considered sample. Although this

approach does not take into account copy number information and assumes similar underlying

cancer cell fraction, we believe that it removes majority of mutations that have ambiguous

placement in phylogetentic trees. Binomial confidence intervals were computed according to

the “bayes” method using binom.confint() function in binom package in R.

Subset analysis to validate the robustness of phylogenetic inference

Intratumor heterogeneity in the primary cancer (11) can complicate the inference of seeding

origin of metastases. In order to ameliorate this effect, we sequenced multiple primary blocks

in some patients which demonstrated, for instance in patient 4, how different primary regions

seeded different metastases (Fig. 3b). However, on the other hand, this also raises the question

whether the number of primary samples sequenced affects the inference of progression model.

This is termed as incomplete taxon sampling problem in phylogenetic inference. In order to

show that Dollo parsimony is robust to this problem, we performed subset analysis for the

following two cases.

i. In case of patient 4, a parallel progression case, where we have 6 primary samples,

taking all 62 possible subsets of primary samples with three metastases and

estimating the probability of linear progression. A case where we observe lower

probability of linear progression in each subset will ultimately support a higher

probability for the existing inference of parallel progression.

ii. In case of patient 5, a linear progression case where we have 2 primary samples,

taking the 2 possible subsets of primary samples with two metastases and

estimating the probability of linear progression. A case where we observe higher

probability of linear progression in each subset will ultimately support the existing

hypothesis reported in the manuscript.

We used the following method to infer the probability of linear progression. We reconstructed

1000 bootstrap trees from available subset of samples as described in the Methods. Then, for

each of the bootstrap tree, we used the separating property to test whether any of the primary

samples is blocking the path among the metastases. If blocking, we have a NO result for

linear progression; if not blocking, we have a YES result for linear progression. Finally we

combined the results across all the 1000 trees to estimate the probability of linear progression.

Subset analysis for patient 4

In patient 4, we have 62 possible subsets. This includes 6 possible subsets containing 1

primary sample, 15 possible subsets containing 2 primary samples, 20 possible subsets

containing 3 primary samples, 15 possible subsets containing 4 primary samples, and 6

possible subsets containing 5 primary samples. The results are given in Supplementary Table

8. We observe from the results that, across all possible 62 subsets, we obtain either zero or

almost zero probability that all three metastases are seeded in a linear fashion. This confirms

that the primary tumor has seeded at least two or all three metastases in parallel, which is in

line with the results for all samples taken together (Fig. 3).

Next, we take into account the paired metastases cases (Uterus to Brain, Uterus to Colon, and

Brain to Colon) where, for a metastases pair, the earlier metastasis has seeded the latter

metastases in a linear fashion. For Uterus to Brain pair, we see only 1/62 case with more than

50% probability meaning that 98% of the subsets support Uterus did not seed the Brain

metastases. For the rest of two possible cases (Uterus-Colon, and Brain-Colon), there is not a

single case with a probability of 50% or higher of linear progression in any pair. Overall, the

subset analysis supports the parallel progression model for patient 4.

Subset analysis for patient 5

In patient 5, we have 2 possible subsets where we take one primary sample each with the two

bone metastases. The results are given in Supplementary Table 9. We observe from the results

that the probability of linear progression in both subsets is almost 100% which supports the

results when full data is used for inferring the progression model.

In summary, we see that Dollo parsimony is robust to the number and combination of primary

samples taken for inferring the phylogenetic tree. This, in turns, means that progression model

inference performed using separating property also does not change when different subsets

are considered.

Parameter values used in PyClone Out of the available three models in PyClone, we used the authors’ recommended genotype-

aware PyClone-beta-binomial model with all model’s parameter values set to recommended

values (the rest of the two models are genotype-naive infinite binomial mixture model and

infinite beta-binomial mixture model). We tested the robustness of cellular prevalence (CP)

cut-off of 0.05 as follows. We set the cellular prevalence cut-off to 0.04 and 0.02 and

compared it to CP cut-off of 0.05 to check if the seeding patterns are altered. We observed

that, overall, the progression and lymph node seeding results were not changed for CP

threshold of 0.04 and they were quite similar for the CP threshold of 0.02 (Supplementary

Table 10). Regarding the number of iterations in MCMC, the following criterion was used. If

the number of samples in patient were less than 5, 10000 iterations were used; if the number

of samples in a patient were between 5 and 7, 15000 iterations were used; if the number of

samples in a patient were between 7 and 10, 20000 iterations were used; and if the number of

samples were more than 10, 50000 iterations were used. The first 25% percent iterations were

thrown as burnin, thereafter every 10th sample was considered, i.e., a thinning value of 10

was used. To test convergence, we ran two independent PyClone analyses for each patient and

compared the results. For patient 11, we found that using 15000 iterations for MCMC

sampling were not enough for convergence. Subsequently, we used 30000 iterations and

observed convergence.

Mutational Signatures

We extracted a number of signatures ranging between 2-10 with five repetitions, and

computed the residuals sum of squares (RSS) and the explained variance between the

observed profile and fitted spectrum for different number of signatures. The final number of

signatures (four) was decided based on the first inflection point when plotting RSS and

explained variance change with number of signatures (Supplementary Fig. 7a). The accuracy

of the fitted signatures is dependent on the number of samples used for extraction. To allow

higher accuracy of fitting, we merged our cohort with an external in-house cohort of primary

breast cancers from 129 patients with exome sequencing. The external cohort analysis was

performed in a similar pipeline, which excludes potential batch effects.

To identify the biological processes underlying each signature, the Euclidean distance was

computed between the frequencies of different mutation classes in our four signatures and

those in the validated signatures published by Alexandrov et al (12). Based on the shortest

Euclidean distance, we were able to reliably map signatures S1 and S2 to the age-associated

signature 1 and APOBEC-associated signature 2 from Alexandrov et al respectively

(Supplementary Fig. 7b). Signature S3 had a similar distance to several published signatures.

We believe that the best candidate for S3 is signature 8 which has an unknown etiology since

they share the characteristic of weak strand bias in C>A substitutions (Supplementary Fig. 7c)

and since signature 8 was also found in breast cancer. We found that elevated contribution of

signature S4 is significantly associated with BRCA1/2 deleterious germline mutations in the

external cohort (p-value = 0.0009, Mann-Whitney). Consequently, S4 was mapped to

signature 3 in Alexandrov et al (12) which is associated with homologous recombination

deficiency.

References1. GuiuS,MichielsS,AndreF,CortesJ,DenkertC,DiLeoA,HennessyBT,SorlieT,

SotiriouC,TurnerN,etal.Molecularsubclassesofbreastcancer:howdowedefinethem?TheIMPAKT2012WorkingGroupStatement.Annalsofoncology:officialjournaloftheEuropeanSocietyforMedicalOncology/ESMO.2012;23(12):2997-3006.

2. CoatesAS,WinerEP,GoldhirschA,GelberRD,GnantM,Piccart-GebhartM,ThurlimannB,SennHJ,andPanelM.Tailoringtherapies--improvingthemanagementofearlybreastcancer:StGallenInternationalExpertConsensusonthePrimaryTherapyofEarlyBreastCancer2015.Annalsofoncology:officialjournaloftheEuropeanSocietyforMedicalOncology/ESMO.2015;26(8):1533-46.

3. PratA,CheangMC,MartinM,ParkerJS,CarrascoE,CaballeroR,TyldesleyS,GelmonK,BernardPS,NielsenTO,etal.Prognosticsignificanceofprogesteronereceptor-positivetumorcellswithinimmunohistochemicallydefinedluminalAbreastcancer.Journalofclinicaloncology:officialjournaloftheAmericanSocietyofClinicalOncology.2013;31(2):203-9.

4. LindahlT,LandbergG,AhlgrenJ,NordgrenH,NorbergT,KlaarS,HolmbergL,andBerghJ.OverexpressionofcyclinEproteinisassociatedwithspecificmutationtypesinthep53geneandpoorsurvivalinhumanbreastcancer.Carcinogenesis.2004;25(3):375-80.

5. SjogrenS,InganasM,NorbergT,LindgrenA,NordgrenH,HolmbergL,andBerghJ.Thep53geneinbreastcancer:prognosticvalueofcomplementaryDNAsequencingversusimmunohistochemistry.JNatlCancerInst.1996;88(3-4):173-82.

6. LinderholmB,KarlssonE,KlaarS,LindahlT,BorgAL,ElmbergerG,andBerghJ.Thrombospondin-1expressioninrelationtop53statusandVEGFexpressioninhumanbreastcancers.Europeanjournalofcancer.2004;40(16):2417-23.

7. StalhammarG,FuentesMartinezN,LippertM,TobinNP,MolholmI,KisL,RosinG,RantalainenM,PedersenL,BerghJ,etal.Digitalimageanalysisoutperformsmanualbiomarkerassessmentinbreastcancer.ModPathol.2016;29(4):318-29.

8. ParkerJS,MullinsM,CheangMC,LeungS,VoducD,VickeryT,DaviesS,FauronC,HeX,HuZ,etal.Supervisedriskpredictorofbreastcancerbasedonintrinsicsubtypes.JClinOncol.2009;27(8):1160-7.

9. ZhaoX,RodlandEA,TibshiraniR,andPlevritisS.Molecularsubtypingforclinicallydefinedbreastcancersubgroups.BreastCancerRes.2015;17(29.

10. PawitanY,BjohleJ,AmlerL,BorgAL,EgyhaziS,HallP,HanX,HolmbergL,HuangF,KlaarS,etal.Geneexpressionprofilingsparesearlybreastcancerpatientsfromadjuvanttherapy:derivedandvalidatedintwopopulation-basedcohorts.BreastCancerRes.2005;7(6):R953-64.

11. YatesLR,GerstungM,KnappskogS,DesmedtC,GundemG,VanLooP,AasT,AlexandrovLB,LarsimontD,DaviesH,etal.Subclonaldiversificationofprimarybreastcancerrevealedbymultiregionsequencing.NatMed.2015;21(7):751-9.

12. AlexandrovLB,Nik-ZainalS,WedgeDC,AparicioSA,BehjatiS,BiankinAV,BignellGR,BolliN,BorgA,Borresen-DaleAL,etal.Signaturesofmutationalprocessesinhumancancer.Nature.2013;500(7463):415-21.

1 2 3 4 5 6 7 8 9 10

1_Pr

imar

y1_

Lung

1.R

11_

Lung

2.R

11_

Live

r1.R

21_

Live

r2.R

21_

Nor

mal

2_Pr

imar

y2_

ALN

2_C

olon

1.R

12_

Col

on2.

R1

2_N

orm

al

3_Pr

imar

y3_

ALN

3_Bo

ne.R

13_

Nor

mal

4_Pr

imar

y64_

Prim

ary4

4_Pr

imar

y24_

Prim

ary3

4_Pr

imar

y14_

Prim

ary5

4_U

teru

s.R

24_

Brai

n.R

34_

Col

on.R

44_

Nor

mal

5_Pr

imar

y15_

Prim

ary2

5_Bo

ne.R

15_

Bone

.R2

5_N

orm

al

6_Pr

imar

y6_

Nor

mal

7_Pr

imar

y.lo

cal.r

ecur

renc

e7_

Skin

1.R

17_

Skin

2.R

17_

Nor

mal

.17_

Nor

mal

.27_

Nor

mal

.3

8_Pr

imar

y8_

ALN

8_Sk

in8_

Bone

8_N

orm

al.1

8_N

orm

al.2

9_Pr

imar

y9_

Brai

n3.R

19_

Brai

n2.R

19_

Brai

n1.R

19_

Nor

mal

.19_

Nor

mal

.2

10_P

rimar

y110

_Prim

ary2

10_A

LN1

10_A

LN2

10_S

kin3

.R2

10_S

kin2

.R2

10_S

kin1

.R2

10_N

orm

al

0.00

0.25

0.50

0.75

1.00

Targ

ets

perc

enta

ge c

over

ed

11 12 13 14 15 16 17 18 19 20

11_P

rimar

y111

_Prim

ary5

11_P

rimar

y311

_Prim

ary4

11_P

rimar

y211

_Bra

in2.

R1

11_B

rain

1.R

111

_Nor

mal

12_P

rimar

y12

_ALN

12_N

orm

al

13_P

rimar

y.lo

cal.r

ecur

renc

e.1

13_P

rimar

y.lo

cal.r

ecur

renc

e.2

13_A

LN.re

laps

e13

_Bon

e1.R

113

_Bon

e2.R

113

_Nor

mal

14_P

rimar

y14

_ALN

14_B

rain

1.R

114

_Bra

in2.

R1

14_N

orm

al

15_P

rimar

y15

_ALN

15_L

iver

.R1

15_N

orm

al.1

15_N

orm

al.2

16_P

rimar

y.lo

cal.r

ecur

renc

e.1

16_P

rimar

y.lo

cal.r

ecur

renc

e.2

16_P

rimar

y.lo

cal.r

ecur

renc

e.3

16_S

kin3

.R1

16_S

kin2

.R1

16_S

kin1

.R1

16_N

orm

al

17_P

rimar

y217

_Prim

ary1

17_A

LN1

17_A

LN2

17_B

rain

.R1

17_N

orm

al

18_P

rimar

y218

_Prim

ary1

18_A

LN1

18_A

LN2

18_S

kin1

.R2

18_S

kin2

.R2

18_N

orm

al

19_P

rimar

y19

_Bra

in1.

R3

19_B

rain

1.R

519

_Bra

in2.

R5

19_N

orm

al

20_P

rimar

y420

_Prim

ary2

20_P

rimar

y120

_Prim

ary3

20_O

vary

8.R

220

_Ova

ry6.

R2

20_O

vary

2.R

220

_Ova

ry5.

R2

20_O

vary

4.R

220

_Ova

ry7.

R2

20_O

vary

9.R

220

_Ova

ry3.

R2

20_O

vary

1.R

220

_Nor

mal

.120

_Nor

mal

.2

sample

Coverage Level> x0

> x5

> x10

> x30

> x50

> x80

sample

0.00

0.25

0.50

0.75

1.00

Targ

ets

perc

enta

ge c

over

ed

Supplementary Figure 1

Prim

ary

Dia

gnos

is

1

2

Neo

Adju

vant

3

4

5

6

7

8

9

10

55 months

1 year

CT

RT

HT

M1 M2

57 monthsM1

L+

11 monthsL+ M1

116 months

M1 M2 M3

M4

58 months

M1

11

12

13

14

15

16

17

18

19

20

M1

92 months

84 months

M2adr.gl

M3boneM1

82 monthsL+ M1

48 months

M2M1

L+M1

76 months

33 months

M1

36 monthsL+M1

163 months

M1

108 monthsL+

142 monthsL+

40 months

M1

80 monthsL+

M1

115 monthsL+

M3

90 months

M1 M2

M1

lung liver

colon

Loc

bone

Contrlung uterus brain

colon

H2

Her2 therapyL+ positive axillary lymph nodeLoc: Local relapseContr: contralateral Grey text: Not available/not sequenced

TAM

PAM

50 p

rimar

y tu

mor

boneM2bone

boneM2bone

M3lung

Locbone

M2liverskin

LA

LB

BL

BL

BL

LocContrskin

M4brain

brain lung

BL

Locskin Contrskin

M2skin

M3bone

BL

brainM2lung

M3skin

Locaxilla bone

Locbreast bone

Locaxillary

BL

M2bone

Contr M2brain

M3esophagusliver

Loc skinM2lung

M3brain

Contr brain

Contr Loc skin skin boneM4lung

M1bone M2

boneM3brain

M4bone

M5brain

Loc

boneM2ovary

295 months

L+

182 months 262 months

LB

H2

LA/LB

LB

5 y 10 y 15 y

LA Luminal ABL Basal like

LB Luminal B

H2 Her 2 enriched

LA/NBL

N/A

N/D

LB

N/A

N/A

N/A

NBL Normal breast like

M1Loc


Exome Sequencing

Data

Somatic Mutations

Copy Number Calls

Tumor cell fraction

Phylogenetic Reconstruction

Subclonal Analysis

100

80

90

PYCLONE

DOLLO PARSIMONY MUTECT

ASCATNGS

Germline

Clu

ster

Clu

ster

Clu

ster

Clu

ster

Clu

ster

Clu

ster

Clu

ster

Clu

ster

Clu

ster

Primary

Primary

one R

one R

Prim

ary

Prim

ary

one

R

one

R


A

B

Primary

Germline

Metastasis 2a Metastasis 2b

Primary

Germline

Lymph node

Metastasis 2a Metastasis 2b

Metastasis 1

Metastasis 1

Supplementary Figure

Liver1.R2

Liver2.R2

Lung1.R1

Lung2.R1

Primary

Primary

Lung2.R1

Lung1.R1

Liver2.R2

Liver1.R2

0.4 0.43 0.46 0.49 1

0.66 0.69 0.79 1 0.49

0.68 0.69 1 0.79 0.46

0.73 1 0.69 0.69 0.43

1 0.73 0.68 0.66 0.4

Percentage of shared mutations in patient 1

Liver1.R2

Liver2.R2

Lung1.R1

Lung2.R1

Primary

Primary

Lung2.R1

Lung1.R1

Liver2.R2

Liver1.R2

0.33 0.29 0.33 0.3 0

0.12 0.1 0.12 0 0.38

0.1 0.1 0 0.12 0.4

0.15 0 0.25 0.26 0.48

0 0.16 0.26 0.28 0.51

Percentage of exclusive (specific) mutation in patient 1

ALN

Colon1.R1

Colon2.R1

Primary

Primary

Colon2.R1

Colon1.R1

ALN

0.44 0.35 0.46 1

0.34 0.4 1 0.46

0.24 1 0.4 0.35

1 0.24 0.34 0.44


ALN

Colon1.R1

Colon2.R1

Primary

Primary

Colon2.R1

Colon1.R1

ALN

0.14 0.57 0.4 0

0.25 0.51 0 0.34

0.26 0 0.31 0.33

0 0.74 0.62 0.52


Percentage of of mutations present in sample on y-axis but absent in sample on x-axis


Supplementary igure 5

ALN

Bone.R1

Primary

Primary

Bone.R1

ALN

0.28 0.4 1

0.23 1 0.4

1 0.23 0.28


ALN

Bone.R1

Primary

Primary

Bone.R1

ALN

0.63 0.35 0

0.72 0 0.48

0 0.47 0.45

Percentage of exclusive (specific) mutation in patient 3Brain.R3

Colon.R4

Primary1

Primary2

Primary3

Primary4

Primary5

Primary6

Uterus.R2

Uterus.R2

Primary6

Primary5

Primary4

Primary3

Primary2

Primary1

Colon.R4

Brain.R3

0.34 0.29 0.32 0.34 0.34 0.36 0.38 0.34 1

0.44 0.4 0.48 0.56 0.61 0.64 0.57 1 0.34

0.42 0.39 0.46 0.56 0.59 0.58 1 0.57 0.38

0.46 0.42 0.5 0.62 0.61 1 0.58 0.64 0.36

0.44 0.4 0.53 0.57 1 0.61 0.59 0.61 0.34

0.42 0.4 0.48 1 0.57 0.62 0.56 0.56 0.34

0.43 0.41 1 0.48 0.53 0.5 0.46 0.48 0.32

0.4 1 0.41 0.4 0.4 0.42 0.39 0.4 0.29

1 0.4 0.43 0.42 0.44 0.46 0.42 0.44 0.34


Brain.R3

Colon.R4

Primary1

Primary2

Primary3

Primary4

Primary5

Primary6

Uterus.R2

Uterus.R2

Primary6

Primary5

Primary4

Primary3

Primary2

Primary1

Colon.R4

Brain.R3

0.44 0.52 0.46 0.45 0.46 0.43 0.41 0.46 0

0.36 0.43 0.32 0.26 0.24 0.2 0.26 0 0.52

0.39 0.45 0.34 0.27 0.26 0.26 0 0.28 0.49

0.36 0.43 0.31 0.23 0.25 0 0.27 0.24 0.51

0.37 0.43 0.27 0.25 0 0.23 0.25 0.24 0.52

0.4 0.44 0.33 0 0.29 0.24 0.29 0.3 0.53

0.4 0.45 0 0.36 0.34 0.35 0.39 0.38 0.56

0.4 0 0.38 0.41 0.42 0.4 0.43 0.43 0.57

0 0.45 0.39 0.42 0.42 0.38 0.43 0.41 0.54





Bone.R1

Bone.R2

Primary1

Primary2

Primary2

Primary1

Bone.R2

Bone.R1

0.59 0.58 0.71 1

0.57 0.6 1 0.71

0.71 1 0.6 0.58

1 0.71 0.57 0.59


Bone.R1

Bone.R2

Primary1

Primary2

Primary2

Primary1

Bone.R2

Bone.R1

0.18 0.2 0.13 0

0.23 0.21 0 0.2

0.16 0 0.28 0.32

0 0.18 0.31 0.32


Primary.local.recurrence

Skin1.R1

Skin2.R1

Skin2.R1

Skin1.R1


0.17 0.19 1

0.44 1 0.19

1 0.44 0.17



Skin1.R1

Skin2.R1

Skin2.R1

Skin1.R1


0.54 0.31 0

0.5 0 0.79

0 0.22 0.78





ALN

Bone

Primary

Skin

Skin

Primary

Bone

ALN

0.41 0.67 0.41 1

0.58 0.38 1 0.41

0.41 1 0.38 0.67

1 0.41 0.58 0.41


ALN

Bone

Primary

Skin

Skin

Primary

Bone

ALN

0.54 0.22 0.51 0

0.32 0.35 0 0.29

0.53 0 0.52 0.16

0 0.26 0.21 0.22


Brain1.R1

Brain2.R1

Brain3.R1

Primary

Primary

Brain3.R1

Brain2.R1

Brain1.R1

0.66 0.65 0.63 1

0.76 0.78 1 0.63

0.8 1 0.78 0.65

1 0.8 0.76 0.66


Brain1.R1

Brain2.R1

Brain3.R1

Primary

Primary

Brain3.R1

Brain2.R1

Brain1.R1

0.1 0.1 0.12 0

0.14 0.12 0 0.31

0.11 0 0.12 0.3

0 0.11 0.13 0.29





ALN1

ALN2

Primary1

Primary2

Skin1.R2

Skin2.R2

Skin3.R2

Skin3.R2

Skin2.R2

Skin1.R2

Primary2

Primary1

ALN2

ALN1

0.15 0.21 0.22 0.2 0.4 0.37 1

0.18 0.2 0.26 0.23 0.37 1 0.37

0.2 0.19 0.26 0.23 1 0.37 0.4

0.27 0.28 0.3 1 0.23 0.23 0.2

0.2 0.27 1 0.3 0.26 0.26 0.22

0.17 1 0.27 0.28 0.19 0.2 0.21

1 0.17 0.2 0.27 0.2 0.18 0.15


ALN1

ALN2

Primary1

Primary2

Skin1.R2

Skin2.R2

Skin3.R2

Skin3.R2

Skin2.R2

Skin1.R2

Primary2

Primary1

ALN2

ALN1

0.79 0.74 0.74 0.7 0.4 0.45 0

0.75 0.75 0.7 0.66 0.45 0 0.48

0.74 0.76 0.7 0.67 0 0.48 0.45

0.62 0.64 0.64 0 0.56 0.58 0.64

0.61 0.56 0 0.36 0.31 0.33 0.44

0.68 0 0.6 0.42 0.5 0.5 0.5

0 0.73 0.71 0.5 0.54 0.58 0.67


Brain1.R1

Brain2.R1

Primary1

Primary2

Primary3

Primary4

Primary5

Primary5

Primary4

Primary3

Primary2

Primary1

Brain2.R1

Brain1.R1

0.14 0.23 0.22 0.23 0.26 0.26 1

0.16 0.28 0.26 0.29 0.3 1 0.26

0.11 0.4 0.26 0.26 1 0.3 0.26

0.21 0.22 0.3 1 0.26 0.29 0.23

0.24 0.21 1 0.3 0.26 0.26 0.22

0.18 1 0.21 0.22 0.4 0.28 0.23

1 0.18 0.24 0.21 0.11 0.16 0.14


Brain1.R1

Brain2.R1

Primary1

Primary2

Primary3

Primary4

Primary5

Primary5

Primary4

Primary3

Primary2

Primary1

Brain2.R1

Brain1.R1

0.78 0.32 0.65 0.61 0.32 0.54 0

0.78 0.32 0.63 0.57 0.34 0 0.63

0.87 0.4 0.71 0.69 0 0.65 0.71

0.71 0.4 0.58 0 0.36 0.53 0.65

0.66 0.36 0 0.51 0.3 0.52 0.64

0.81 0 0.77 0.74 0.46 0.68 0.74

0 0.32 0.56 0.56 0.59 0.62 0.71




J


ALN.relapse

Bone1.R1

Bone2.R1

Primary.local.recurrence.1




Bone2.R1

Bone1.R1

ALN.relapse

0.37 0.31 0.29 0.37 1

0.31 0.27 0.22 1 0.37

0.48 0.62 1 0.22 0.29

0.5 1 0.62 0.27 0.31

1 0.5 0.48 0.31 0.37


ALN.relapse

Bone1.R1

Bone2.R1





Bone2.R1

Bone1.R1

ALN.relapse

0.38 0.47 0.56 0.43 0

0.49 0.56 0.66 0 0.49

0.24 0.14 0 0.61 0.54

0.31 0 0.31 0.59 0.56

0 0.35 0.43 0.56 0.52


ALN

Brain1.R1

Brain2.R1

Primary

Primary

Brain2.R1

Brain1.R1

ALN

0.24 0.59 0.55 1

0.2 0.64 1 0.55

0.22 1 0.64 0.59

1 0.22 0.2 0.24


ALN

Brain1.R1

Brain2.R1

Primary

Primary

Brain2.R1

Brain1.R1

ALN

0.63 0.2 0.23 0

0.71 0.22 0 0.35

0.67 0 0.21 0.31

0 0.59 0.62 0.6




K


ALN

Liver.R

1

Primary

Primary

Liver.R1

ALN

0.29 0.12 1

0.1 1 0.12

1 0.1 0.29


ALN

Liver.R

1

Primary

Primary

Liver.R1

ALN

0.6 0.65 0

0.88 0 0.85

0 0.65 0.49





Skin1.R1

Skin2.R1

Skin3.R1

Skin3.R1

Skin2.R1

Skin1.R1




0.08 0.09 0.08 0.45 0.44 1

0.12 0.12 0.13 0.43 1 0.44

0.09 0.1 0.08 1 0.43 0.45

0.4 0.45 1 0.08 0.13 0.08

0.41 1 0.45 0.1 0.12 0.09

1 0.41 0.4 0.09 0.12 0.08





Skin1.R1

Skin2.R1

Skin3.R1

Skin3.R1

Skin2.R1

Skin1.R1




0.68 0.55 0.58 0.43 0.32 0

0.57 0.49 0.45 0.49 0 0.44

0.57 0.42 0.5 0 0.27 0.32

0.49 0.39 0 0.91 0.86 0.91

0.48 0 0.37 0.89 0.86 0.9

0 0.35 0.34 0.9 0.86 0.91





ALN1

ALN2

Brain.R1

Primary1

Primary2

Primary2

Primary1

Brain.R1

ALN2

ALN1

0.25 0.27 0.04 0.26 1

0.22 0.22 0.04 1 0.26

0.04 0.04 1 0.04 0.04

0.52 1 0.04 0.22 0.27

1 0.52 0.04 0.22 0.25


ALN1

ALN2

Brain.R1

Primary1

Primary2

Primary2

Primary1

Brain.R1

ALN2

ALN1

0.61 0.62 0.86 0.63 0

0.61 0.65 0.87 0 0.54

0.95 0.95 0 0.96 0.94

0.24 0 0.84 0.64 0.52

0 0.37 0.88 0.68 0.6


ALN1

ALN2

Primary1

Primary2

Skin1.R2

Skin2.R2

Skin2.R2

Skin1.R2

Primary2

Primary1

ALN2

ALN1

0.26 0.14 0.2 0.26 0.73 1

0.28 0.15 0.21 0.26 1 0.73

0.25 0.35 0.3 1 0.26 0.26

0.26 0.32 1 0.3 0.21 0.2

0.23 1 0.32 0.35 0.15 0.14

1 0.23 0.26 0.25 0.28 0.26


ALN1

ALN2

Primary1

Primary2

Skin1.R2

Skin2.R2

Skin2.R2

Skin1.R2

Primary2

Primary1

ALN2

ALN1

0.61 0.64 0.65 0.39 0.14 0

0.6 0.62 0.64 0.4 0 0.17

0.7 0.47 0.64 0 0.68 0.68

0.62 0.31 0 0.35 0.66 0.67

0.73 0 0.63 0.5 0.81 0.82

0 0.4 0.55 0.37 0.54 0.57





Brain1.R3

Brain1.R5

Brain2.R5

Primary

Primary

Brain2.R5

Brain1.R5

Brain1.R3

0.12 0.08 0.11 1

0.67 0.48 1 0.11

0.4 1 0.48 0.08

1 0.4 0.67 0.12


Brain1.R3

Brain1.R5

Brain2.R5

Primary

Primary

Brain2.R5

Brain1.R5

Brain1.R3

0.71 0.67 0.7 0

0.27 0.11 0 0.85

0.58 0 0.49 0.91

0 0.09 0.1 0.83


Ovary1.R2

Ovary2.R2

Ovary3.R2

Ovary4.R2

Ovary5.R2

Ovary6.R2

Ovary7.R2

Ovary8.R2

Ovary9.R2

Primary1

Primary2

Primary3

Primary4

Primary4

Primary3

Primary2

Primary1

Ovary9.R2

Ovary8.R2

Ovary7.R2

Ovary6.R2

Ovary5.R2

Ovary4.R2

Ovary3.R2

Ovary2.R2

Ovary1.R2

0.04 0.15 0.17 0.16 0.16 0.16 0.16 0.18 0.17 0.27 0.31 0.29 1

0.05 0.18 0.18 0.2 0.19 0.19 0.19 0.21 0.19 0.33 0.35 1 0.29

0.06 0.2 0.22 0.22 0.21 0.2 0.22 0.22 0.24 0.36 1 0.35 0.31

0.04 0.17 0.19 0.19 0.2 0.18 0.19 0.2 0.2 1 0.36 0.33 0.27

0.22 0.52 0.52 0.51 0.59 0.5 0.55 0.58 1 0.2 0.24 0.19 0.17

0.24 0.53 0.57 0.55 0.59 0.54 0.6 1 0.58 0.2 0.22 0.21 0.18

0.22 0.54 0.6 0.55 0.61 0.56 1 0.6 0.55 0.19 0.22 0.19 0.16

0.23 0.51 0.56 0.56 0.55 1 0.56 0.54 0.5 0.18 0.2 0.19 0.16

0.25 0.62 0.57 0.55 1 0.55 0.61 0.59 0.59 0.2 0.21 0.19 0.16

0.21 0.52 0.55 1 0.55 0.56 0.55 0.55 0.51 0.19 0.22 0.2 0.16

0.21 0.52 1 0.55 0.57 0.56 0.6 0.57 0.52 0.19 0.22 0.18 0.17

0.25 1 0.52 0.52 0.62 0.51 0.54 0.53 0.52 0.17 0.2 0.18 0.15

1 0.25 0.21 0.21 0.25 0.23 0.22 0.24 0.22 0.04 0.06 0.05 0.04


Ovary1.R2

Ovary2.R2

Ovary3.R2

Ovary4.R2

Ovary5.R2

Ovary6.R2

Ovary7.R2

Ovary8.R2

Ovary9.R2

Primary1

Primary2

Primary3

Primary4

Primary4

Primary3

Primary2

Primary1

Ovary9.R2

Ovary8.R2

Ovary7.R2

Ovary6.R2

Ovary5.R2

Ovary4.R2

Ovary3.R2

Ovary2.R2

Ovary1.R2

0.95 0.8 0.75 0.77 0.78 0.76 0.78 0.75 0.77 0.64 0.59 0.61 0

0.94 0.74 0.7 0.68 0.71 0.68 0.7 0.66 0.71 0.51 0.48 0 0.45

0.92 0.72 0.66 0.65 0.7 0.66 0.66 0.66 0.66 0.48 0 0.49 0.44

0.94 0.74 0.68 0.69 0.71 0.69 0.69 0.67 0.7 0 0.45 0.49 0.48

0.75 0.34 0.25 0.26 0.28 0.25 0.24 0.22 0 0.64 0.57 0.64 0.6

0.74 0.37 0.26 0.28 0.32 0.26 0.25 0 0.31 0.65 0.62 0.64 0.62

0.76 0.36 0.24 0.28 0.3 0.25 0 0.25 0.33 0.67 0.62 0.67 0.66

0.76 0.41 0.3 0.3 0.37 0 0.31 0.32 0.4 0.7 0.66 0.69 0.67

0.72 0.25 0.19 0.21 0 0.19 0.17 0.19 0.24 0.63 0.6 0.63 0.61

0.77 0.39 0.29 0 0.36 0.26 0.3 0.3 0.37 0.68 0.63 0.66 0.66

0.77 0.39 0 0.29 0.34 0.27 0.27 0.29 0.37 0.68 0.63 0.69 0.64

0.72 0 0.22 0.23 0.23 0.21 0.22 0.23 0.29 0.67 0.61 0.65 0.63

0 0.31 0.29 0.29 0.29 0.22 0.29 0.24 0.33 0.82 0.74 0.8 0.78





Patient 1: ER-/PR-/HER2-

Cluster color

Cluster ID Mutation count

Driver genes

Truncal 1 99 TP53;NCOA3

17 39 0 Primary 6 47 MAP3K6;KDM5C Metastasis 7 81 XPO1;MED12

12 29 0

8 33 0

CDK6

100

100

76

100Primary Lung2.R1

Lung1.R1

Liver2.R2 Liver1.R2

TP53NCOA3AKT3MYCHRASRB1STK11(102)

MAP3K6KDM5C

(77)

(38)

(29)

MED12(59)

(10)

XPO1

(36)

(37)

(36)

(19 mo)

(19 mo)

(49 mo) (49 mo)

ERBB2MCL1MDM4RELCCNE1MDM2AKT1IGF1RPIK3CAPDGFRATERTEGFRBRCA1CDKN1BPBRM1PTEN

CCNE1FGFR3ARID1ABAP1PBRM1

ERBB2

MCL1MDM4MYCNRELARID1ACDKN1BBAP1PBRM1

Germline

A

Primary

Lung2.R1

Liver2.R2

Liver1.R2

Phylogenetic Tree

Cluster Table

Density Plot

u le entary igure


Cluster color


Driver genes

Truncal

1 21 TP53

3 60 BAP1

7 190 ERBB3;NF1;EPHA5;RRM2B

10 (except lymph) 31 0 Primary & Lymph

6 10 0

Lymph 2 193 KAT6B;MED12Metastasis 9 53 0

Colon2.R1

Germline

Primary

ALN

Colon1.R1

(78)

BAP1RRM2B

(84)

ERBB3(89)

KAT6BERBB3MED12MCL1PDGFRAMYC

(233)

(61)

RRM2B

(75)

(36)

100

100

100

TP53

ERBB3

ZNF217AURKA

BAP1EPHAS

NF1

(57 mo)

(57 mo)

NOTCH3CCNE1AKT2PIK3CA

B

Primary

ALN

Colon2.R1

Colon1.R1

Phylogenetic Tree

Cluster Table

Density Plot

u le entary igure

Patient 3: ER-/PR-/HER2+

Cluster color


Driver genes

Truncal 1 13 0 Primary & Lymph

2 59 FANCA

Primary & Metastasis

3 25 PIK3CA;FGFR4

Lymph & Metastasis

4 55

100

100

Primary

ALN

Bone.R1

(20)

FANCA(33)

PIK3CAFGFR4TERTMYC

(32)

(47)(29)

Germline

(11 mo)

ERBB2

AKT3CCND1MDM2NSD1ARID1ARB1BRCA1BAP1PBRM1CDKN2ACDKN2B

0

C

Primary

ALN

Bone.R1

Phylogenetic Tree

Cluster Table

Density Plot

u le entary igure

D

Colon.R4

Primary

1

Uterus.R2

Primary5Prim

ary2

Primary

4Primary

3

Primary6

Brain.R3

Germline

JAK1

ERBB2

(58)

DNMT3A

BRCA2

IGF1R

ZNF217

AURKA

NOTCH3

CCNE1

BRCA1

BAP1

PBRM1

CDKN2A

CDKN2B

(103)

DDX11

(3)(2)

(0)

(115)PBRM1

MDM4

AKT3

IGF1R

MYCN

ZNF217

AURKA

TERT

MYC

ARID1A

RB1

BRCA1

STK11

BAP1

PBRM1

CDKN2A

CDK2B

HRAS

(118)

(55) (38)

(42)

(11)

DDR2

BRCA2

(60)

(79)

(17)

(94)

(7)

(54)

100

100

26 79

39

42

6473

(74 mo)

(93 mo)

EPHA3

ARID1A

RB1

BRCA1

STK11

BAP1

PBRM1

CDKN2A

CDKN2B

HRAS

PTEN

Cluster color


Driver genes

Truncal

20 15

18 47 DNMT3A

3 70 JAK1

12 (except Brain.R3) 29 0

Primary

1 / P6,4,2,3,5 15 0

6 / P1 44 PBRM1

8 / P2 27 EPHA3

15 / P4 47 PLCG1

9 / P6 24 0

Primary 16 & Uterus.R2

14 / P5, Uterus.R2 79 BRCA2;DDR2;ROS1;KDM6A

Metastasis 4 / Colon 45 BRCA2

5 / Brain 57

0

0

AKT3

TERT


(97 mo)

AKT1

IGF1R

PAX8

MYCN

NSD1

NOTCH3

CCNE1

AKT2

EGFRDNMT3A

FGFR2

FGFR3

KDM6A

MDM4

ZNF217

AURKA

ARID1A

RB1

BRCA1

STK11

BAP1

PBRM1

CDKN2A

CDKN2B

HRAS

DNMT3A

ROS1

MDM4

AKT3

AKT1

IGF1R

PAX8

MYCN

TERT

NSD1

NOTCH3

CCNE1

EGFR

CDK6

MYC

ARID1A

RB1

BRCA1

STK11

BAP1

PBRM1

CDKN2A

CDKN2B

DNMT3A

MDM4

AKT3

AKT1

PAX8

TERT

NSD1

NOTCH3

CCNE1

AKT2

EGFR

CDK6

ARID1A

RB1

BRCA1

STK11

BAP1

PBRM1

CDKN2A

CDKN2B

HRAS

MDM4

AKT3AKT1

IGF1R

PAX8

MYCN

ZNF217

AURKA

TERT

NSD1

ARID1A

RB1

BRCA1

STK11

BAP1

PBRM1

CDKN2A

CDKN2B

Uterus.R2

Primary5

Primary1

Primary3

Primary2

Primary4

Primary6

Colon.R4

Brain.R3

Phylogenetic Tree

Cluster Table

Density Plot

u le entary igure

Patient 5: ER+/PR+/HER2-

Cluster color


Driver genes

Truncal 2 12 0

3 121 GATA1

9 30 PPM1D

Primary 8 10 0 11 10 0 12 18

Metastasis 16 34 0

1 18 0

15 24 0

0

10066

100

Primary1

Primary2Bone.R2 Bone.R1

(124)PPM1DCCND1IGF1RZNF217AURKAERLIN2TERTARID1APTENBRCA1

(51)

(53)

(10)

(38)

PPM1DMDM4CCND1IGF1RZNF217AURKAESR1(38)

(32)

GATA1CDK4MDM2PIK3CA

MST1SLIT2CCND1ZNF217AURKAERLIN2CCNE1ARID1APTENBRCA1BAP1PBRM1CDKN2ACDKN2BCDKN1B

Germline

(49 mo) (31 mo)

MCL1

ERLIN2ARID1APTENBRCA1BAP1PBRM1CDKN2ACDKN2B

E

Primary1

Primary2

Bone.R2

Bone.R1

Phylogenetic Tree

Cluster Table

Density Plot

u le entary igure

Patient 7: ER+/PR-/HER2- (Skin Metastasis IHC, Primary Tumor data NA)

Cluster color


Driver genes

Truncal 9 38 TP53

28 134 Primary 2 16 0

0

100

100

Primary local recurrence

Skin1.R1

Skin2.R1

(21)

(80)

(29)

(103)

(41)

Germline

(64 mo)

(64 mo)

MYC

MCL1NKX2-1FGFR3PDGFRABAP1PBRM1CDKN2ACDKN2B

CCND1ZNF217

(54 mo)

F

Primary local recurrence

Skin2.R1

Skin1.R1

Phylogenetic Tree

Cluster Table

Density Plot

u le entary igure

Patient 8: ER+/PR+/HER2- (Based on IHC on axillary Lymph and Bone metastasis)

Cluster color


Driver genes

Truncal 6 86 0Primary 14 41 0Lymph 15 18 0Metastasis 4 57 EP400;STAT3

2 40 PLCG1

1 27 NOTCH3

ALN

100

100

56

(82)EP400STAT3AKT3MDM2NKX2-1MYCNCDKN2ACDKN2B

(71)

NOTCH3NOTCH3MDM4CCND1IGF1RPIK3CATERTPTENRB1(29)

Primary

Local regionalSkin.R1

Bone.R2(41) (23)

PLCG1CDKN1B

(44)

(13)

Germline

(49 mo)

(75 mo)

ERBB2ZNF217AURKAMYCHRASSTK11

MDM4CCND1IGF1R

AKT3MDM2TERTCDK4NOTCH3AKT2NFE2L2EGFRFGFR1PTENRB1CDKN1B

MYCN

G

Primary

ALN

Bone.R2

Local regional

Skin.R1

Phylogenetic Tree

Cluster Table

Density Plot

u le entary igure


Germline

(26 mo)

(26 mo)(26 mo)

100

100

84Primary

Brain1.R1

Brain2.R1Brain3.R1

HSP90AB1KRASNKX2-1NOTCH3ZNF217AURKACCND3MYCCDKN1BBAP1PBRM1

(262)

NTRK3PTEN

(63)

(45)

(19)

(48)

MST1AKT3(46)

(62)

PPM1DNF2

MCL1IGF1RHRAS

MCL1IGF1RNSD1

Cluster color


Driver genes

Truncal 2 166 0

8 74 HSP90AB1

22 37 0

Primary 1 31 0

Metastasis

30 53 PPM1D;NF2

29 34 NTRK3

16 23 MST1

17 13 0

19 51 0

H

Primary

Brain1.R1

Brain2.R1

Brain3.R1

Phylogenetic Tree

Cluster Table

Density Plot

u le entary igure


100

100

71

59

4438

Primary1

Primary2

ALN2

Skin1.R2Skin2.R2

Skin3.R2

ALN1

INSRFGFR2CCND1CDK4MDM2

(59)

(0)

(22)

(9)(27)

DDX3XCDK4MCL1PDGFRAEGFRCDK6ERLIN2WHSC1L1FGFR1

(48)

FANCD2

PHLPP2

(39)

(2) (5)

(30)

(6)

(34)

(43)

CCND1

FGFR2ESR1MYC1BRCA1

FGFR2

BRCA1

Germ

line

(54 mo)

(54 mo)

(54 mo)

Cluster color


Driver genes

Truncal 8 20 0

9 (except primary1)

37 FANCD2

Primary 13 55 INSR

Metastasis

2 11 0

12 36 DDX3X

7 27 PHLPP2

6 19 0

I

Primary1

Primary2

ALN2

Skin3.R2

Skin2.R2

Skin1.R2

Phylogenetic Tree Cluster Table

Density Plot

u le entary igure


(24 mo)

(24 mo)

Germline

Primary1

Primary2

Primary3

Primary4

Primary5

Brain2.R1

Brain1.R1

CARD11TP53MYC

(40)

(255)

(719)

NOTCH2DDR2MST1ARID1BMCL1AKT3CCNE1AKT2CCND3MDM4AKT1IGF1RNOTCH3NFE2L2TERTEGFRPTENHRASBRCA1BAP1PBRM1

(187)

(616)

(241)

BCL2L1PTENHRASBRCA1STK11BAP1PBRM1

(164)

(19)

(294)

(4)

(12)

(32)

(36)

100

100

55

56

42

59

MST1CDK6MCL1AKT3MDM4TERTPAX8BRCA1STK11

NOTCH2DDR2MST1ARID1BMCL1AKT3CCNE1AKT2CCND3PTENHRASRB1BRCA1STK11BAP1PBRM1

MCL1AKT3AKT1IGF1RTERTPAX8PTENHRASRB1BRCA1STK11BAP1PBRM1

MCL1

CCNE1TERT

MCL1AKT3AKT2MDM4AKT1IGF1RPAX8KRAS

Cluster color


Driver genes

Truncal

3 32 CARD11

9 36 0

24 (expect Brain1.R1)

484 0

Primary 26 26 CDK6 Primary & Metastasis

22 471 BCL2L1

Metastasis 19 67 NOTCH2;DDR2

16 27 0

J

Brain2.R1

Brain1.R1

Primary1

Primary5

Primary3

Primary4

Primary2

Phylogenetic Tree

Cluster Table

Density Plot

u le entary igure

Patient 13: ER+/PR+/HER2- (Based on Bone Metastasis IHC, Primary Tumor data NA)

Germline

Primary local recurrence 1

Local regional axillary Lymph relapse

Bone2.R1

Bone1.R1

(51) (20)

(151)

(112)

(110)

TSC1(24)

(69)ESR1FLT1(54)

(25)

100

7873

78


EGFRBAP1

(84 mo)

(84 mo)

(≈ 36 mo)

(156 mo)

(156 mo)

Cluster color


Driver genes

Truncal 2 52 0

1 14 0

13 128 0

Primary local reccurence

11 59 NF110 36 0

Lymph relapse & Metastasis

7 56 0

8 65 FLT1;ESR1;TSC1

MDM4ERBB2MYCERLIN2WHSC1L1FGFR1

EGFRKRASIGF1RNOTCH3CCNE1CDK6BAP1

MCL1

MCL1

STK11BAP1

K


Primary local

recurrence 2

Local regional axillary Lymph relapse

Bone2.R1

Bone1.R1


Density Plot

u le entary igure

Patient 14: ER+/PR+/HER2+

Cluster color


Driver genes

Truncal

2 42 0

13 45 0

7 (except lymph) 94 0Primary 10 24 0Lymph 6 85 CCNE1 Metastasis 11 63 0

L

Primary

ALN

Brain2.R1

Brain1.R1

Phylogenetic Tree

Cluster Table

Density Plot

Germline

Primary

Brain1.R1 Brain2.R1

ALN

(49)

(41)

(76)

(50)

(109)

(33)

(47)

100

100

100

(26 mo)(26 mo)

AKT3ERBB2TERT

NOTCH3BRCA1RB1STK11CDKN2ACDKN2B

MDM4CCND1KRASZNF217AURKA

EGFRPAX8

MYCBRCA1

EGFRBRCA1

HRASBAP1

u le entary igure


Cluster color


Driver genes

Truncal 5 27 0 Primary 9 62 IGF2R

Metastasis 18 137 TP53;IKBKE

20 49 0

M

Primary

Liver.R1

Primary

Germline

ALN

Liver.R1

(20)

(20)

IKBKETP53MDM4AKT3TERTEGFRARID1APTENCDKN1BRB1BRCA1STK11

(202)

IGF2RERBB2

(57)

MAP3K4

(36)

77

100

(116 mo)

ERBB2

C

D

F

Germline

Skin1.R1

Skin2.R1

Skin3.R1




TP53ERBB2ERLIN2

(23)WHSC1L1AKT1BRCA1

(86)

(146)

(218)

WHSC1L1

(126)

WHSC1L1FGFR3AKT3AKT1ERBB2TERTEGFRMYCHRASCDKN1BBRCA1STK11CDKN2ACDKN2B

(93)

(2)

(6)

RPS6KB1PAK3

(172)

(239)

WHSC1L1FGFR3

(200)

88

100100

100

100

RPS6KB1

FGFR3

(16 mo)

(16 mo)

(16 mo)


Cluster color


Driver genes

Truncal 11 51 TP53

6 (except Skin3) 45 RPS6KB1Primary local recurrence 1 727 PAK3

Metastasis 14 50 RPS6KB1

(12 mo)

(12 mo)

(12 mo)

N

Skin3.R1

Skin2.R1

Skin1.R1



Primary local

recurrence 1

C

D

F


Germline

Primary1

Primary2

ALN2

ALN1

Brain.R1

(244)

ARID1AMCL1DDR2PIK3CAKDM6AAKT3AKT1NKX2-1CCNE1NFE2L2ZNF217AURKACCND3CDK6JAK2CDKN1BBRCA1

(22)

(59)

(1)

DDX3XATRXNSD1ERLIN2WHSC1L1FGR1

(71)(9)

ARID1A(31)

PLCG1NSD1HRASRB1CDKN2ACDKN2B

(36)

(19)

100100

46

100

(75 mo)

Cluster color


Driver genes

Truncal 12 27 0

2 (except Brain) 27 0 Primary 9 48 0

5 53 DDR2;DDX3X;ATRXLymph 8 52 ARID1A;PLCG1 Metastasis 4 217 MCL1;DDR2;ARID1A;PIK3CA;

KDM6A

MCL1MDM4

CCND1MYCERLIN2

WHSC1L1FGFR1

AKT3

O

Primary1

Primary2

ALN2

ALN1

Brain.R1

Phylogenetic Tree

Cluster TableDensity Plot

u le entary igure


ABL2TET2FGFR3MYCERLIN2WHSC1L1FGFR1AKT3MCL1MDM4TERT

(62)

PIK3R1ERLIN2AKT3RB1

(94)

NOTCH2PTENRB1ARID1ACDKN2ACDKN2B

(19)

(15)(61)

(55)

PTENATRX

(28) (0)

JAK1NOTCH2ERBB2FGFR3FGFR2CCNE1AKT2MYCERLIN2WHSC1L1FGFR1PTENSTK11RB1

(67)

MAP3K9ERBB2FGFR3(107)

(25)

Germline

ALN1

ALN2Primary2

Primary1

Skin1.R2 Skin2.R2

86

100100

99

78MAP3K9INSR

(87 mo)(87 mo)

Cluster color


Driver genes

Truncal 3 33 PTEN;ATRXPrimary 15 56 INSR;PIK3R1Primary & Lymph 2 14 197 MAP3K9Lymph1 13 23 JAK1 Skin Metastasis & Primary 2

10 63 ABL2;TET2

STK11

PTEN

IGF1RZNF217AURKAMITFNSD1

P

Primary1

Primary2

ALN1

ALN2

Skin1.R2

Skin2.R2

Phylogenetic Tree

Cluster TableDensity Plot

u le entary igure

Patient 19: ER+/PR+/HER2+

100

100

100

Primary

Brain.R3

Brain1.R5

Brain2.R5

(23)

GNASSTK11

(71)

(36)

(22)

PALB2MCL1MDM4AKT2TERTMYCARID1APTENHRASCDKN1BBRCA1BAP1PBRM1CDKN2ACDKN2B

(113)

(20)

(170)

TP53ERBB2

CCND3AKT3ZNF217AURKAJUNNFE2L2

AKT3ZNF217AURKASTK11

Germline

(88 mo)

Cluster color


Driver genes

Truncal 36 20 0 Primary 40 75 GNASMetastasis 9 100 PALB2;TP53

6 12 0

28 18 0

20 27 0

18 203 0

(73 mo)

(88 mo)

Q

Primary

Brain1.R5

Brain2.R5

Brain.R3


Density Plot

u le entary igure

(50)SF3B1MDM4CCND1TERTAKT3

(103)

SF3B1MDM4CCND1IGF1RERLIN2MCL1TERTRB1CDKN2ACDKN2B

(83)

(47)

MAP3K6(48)SF3B1

MDM4CCND1IGF1RERLIN2MCL1AKT1TERTRB1CDKN2ACDKN2B

(66)

(34) (19)

(37)

(51)

(1) (22) (6) (2) (5) SF3B1MCL1(13)

(11)

(34)(72)

CICKDM5CMDM4CCND1IGF1RNKX2-1RB1

(142)

(110)

CICCCND1

(114)

(71)

(29)

(45)

Germ

line

Ovary1.R2

Ovary2.R2

Ovary4.R2

Ovary3.R2Ova

ry5.R2

Ovary6

.R2

Ovary7

.R2 Ovary8.R

2

Ovary9.R

2 Primary1

Primary2 Primary4

Primary3

100

63

2267

49

100 16 6

22 10053

37

SF3B1IGF1RERLIN2MYCN

EGFRMYC

CDKN1B


Cluster color


Driver genes

Truncal

8 19 0

1 (except ovary1.R2) 43 SF3B1

17 (except primary1) 17 0 Primary 27 316 CIC; KDM5C

Primary & Metastasis

14 (In primary3 & all ovary metastasis)

31 0

Metastasis 5 (except ovary1.R2) 30 0

20 28 0

21 26 0

24 24 MAP3K6

23 24 0

29 36 0

KRAS

CCND1NKX2-1RB1CDKN2ACDKN2B

CDK4CDKN2ACDKN2B

MDM4CCND1RB1CDKN2ACDKN2B

R

Primary3

Primary1

Primary2

Primary4

Ovary1.R2

Ovary3.R2

Ovary9.R2

Ovary7.R2

Ovary4.R2

Ovary2.R2

Ovary6.R2

Ovary8.R2

Phylogenetic Tree

Cluster Table

Density Plot

u le entary igure

0

2

4

2 4 6 8 10Number of Signatures

RSS

0.0

0.2

0.4

0.6

2 4 6 8 10Number of Signatures

Expl

aine

dVar

ianc

e

A

S1 S2 S3 S4

S1AS1BS2S3S4S5S6S7S8S9S10S11S12S13S14S15S16S17S18S19S20S21SR1SR2SR3SU1SU2

0.1

0.2

0.3

0.4

0.5 S1

S2

S3

S4

T>GT>CT>AC>TC>GC>A

T>GT>CT>AC>TC>GC>A

T>GT>CT>AC>TC>GC>A

T>GT>CT>AC>TC>GC>A

0.0

0.1

0.2

0.3

value

subs

titut

ion

strandUntranscribed

Transcribed

B







GL

74

52

47

8 2

16

Lung2.R1

0.09

Lung1.R1

0.17 20 29

Liver2.R2

0.11

Liver1.R2

0.13

Primary

0.17


GL

133

36

Colon1.R1

0.0848

208

Primary

0.11

ALN

0.11

Colon2.R1

0.13


GL

14

38

14

21 33

Primary

0.1

Bone.R1

0.12

ALN

0.16


GL

67

18

19 26

36

22

50

23

46 20

46

Brain.R3

0.07

Primary3

0.08

Primary5

0.08

Uterus.R2

0.11

Primary4

0.07

Primary1

0.16

Primary2

0.15

Colon.R4

0.17

Primary6

0.12


GL

137

14 36

9 19

Primary2

0.13

Primary1

0.12

21 17

Bone.R2

0.09

Bone.R1

0.14


GL

19 73

90

Skin2.R1

0.09 15

Skin1.R1

0.15


0.140.14


GL

65

40 17

53

Primary

0.08

ALN

0.13 27 36

Bone

0.26

Skin

0.11


GL

247

30

50

Primary

0.09

14

30

41 22

Brain3.R1

0.1

Brain2.R1

0.11

Brain1.R1

0.11


GL

1118

3820 28133 20 6

Skin3.R2

0.1

Skin2.R2

0.12

Skin1.R2

0.1

ALN1

0.14

ALN2

0.17

Primary2

0.1

Primary1

0.09 0.110.19 0.1


GL

53

31

32

Primary1

0.5

8844

34

64

65

Primary4

0.11

Brain2.R1

0.11

Primary5

0.17

Primary3

0.09

Primary2

0.130

199

0.14

Brain1.R1

0.21

0.141080.12

0.14


GL

121

94 137

ALN

0.14

Primary

0.14


GL

53

27

12

68

34

17


0.1 16 49

12

23

Bone2.R1

0.11

Bone1.R1

0.11 0.09


0.11

ALN.relapse

0.13

0.13 0.12


GL

25

101

84

25

24

Primary

0.11 18 29

ALN

0.15

Brain1.R1

0.1

Brain2.R1

0.13


GL

176

24

Liver.R1

0.3 26440.1

ALN

0.18

Primary

0.11


GL

268

Skin3.R1

0.5

Skin2.R1

0.5

Skin1.R1

0.5149


0.11

75


0.13


0.16


GL

203

23

Brain.R1

0.3

35

56 38

15 5

ALN1

0.14

ALN2

0.12

Primary2

0.11

Primary1

0.1


GL

16

20

25 20

ALN1

0.15

15 47 34

Primary2

0.13

12

ALN2

0.09

Primary1

0.18

53

Skin2.R2

0.13

Skin1.R2

0.18

73 34

0.13 0.16


GL

20

65

111

31

8 209

Brain2.R5

0.08

Brain1.R5

0.08

Primary

0.11 10

Brain1.R3

0.07


GL

27

Ovary1.R2

0.5

21

25

41 27

17

Primary1

0.11

Primary3

0.14 23 88 22 22 19 25 24

Ovary2.R2

0.09

Ovary5.R2

0.12

Ovary9.R2

0.11

Primary2

0.11

Primary4

0.12

Ovary3.R2

0.23

Ovary6.R2

0.19

Ovary8.R2

0.13

Ovary4.R2

0.12

Ovary7.R2

0.16


Mutation map for Patient 1PrimaryLung1.R1Lung2.R1Liver1.R2Liver2.R2


Mutation map for Patient 2PrimaryALNColon1.R1Colon2.R1


Mutation map for Patient 3PrimaryALNBone.R1


Mutation map for Patient 4Primary1Primary2Primary3Primary4Primary5Primary6Uterus.R2Brain.R3Colon.R4


Mutation map for Patient 5Primary1Primary2Bone.R1Bone.R2


Mutation map for Patient 7Primary.local.recurrenceSkin1.R1Skin2.R1


Mutation map for Patient 8PrimaryALNBoneSkin


Mutation map for Patient 9PrimaryBrain1.R1Brain2.R1Brain3.R1


Mutation map for Patient 10Primary1Primary2ALN1ALN2Skin1.R2Skin2.R2Skin3.R2


Mutation map for Patient 11Primary1Primary2Primary3Primary4Primary5Brain1.R1Brain2.R1


Mutation map for Patient 13Primary.local.recurrence.1Primary.local.recurrence.2ALN.relapseBone1.R1Bone2.R1


Mutation map for Patient 14PrimaryALNBrain1.R1Brain2.R1


Mutation map for Patient 15PrimaryALNLiver.R1


Mutation map for Patient 16Primary.local.recurrence.1Primary.local.recurrence.2Primary.local.recurrence.3Skin1.R1Skin2.R1Skin3.R1


Mutation map for Patient 17Primary1Primary2ALN1ALN2Brain.R1


Mutation map for Patient 18Primary1Primary2ALN1ALN2Skin1.R2Skin2.R2


Mutation map for Patient 19PrimaryBrain1.R3Brain1.R5Brain2.R5


Mutation map for Patient 20Primary1Primary2Primary3Primary4Ovary1.R2Ovary2.R2Ovary3.R2Ovary4.R2Ovary5.R2Ovary6.R2Ovary7.R2Ovary8.R2Ovary9.R2


Validationofphylogenetictreesusingmoreconservativemutation-filteringcriteria

Forthephylogenetictreesreportedinthemanuscript(Figure2,3,4andSupplementaryFigure6),weusedmutationsobtainedusing

mutation-callingcriteriaasdescribedintheMethodssection.Next,wetestedtherobustnessoftheseresultsbyremovingmutations

affectedbyvariablecoverageand/ordifferenttumorpurityamongsamples.Forthis,weusedmutation-filteringcriteriaasdescribedin

“Validationofphylogenetictrees”subsectionoftheMethodsection.

Inthefollowingpages,wepresent,foreachpatient,aside-by-sidecomparisonofthetreereportedinthemanuscript(termedhereOldtree)vs.theonereconstructedusingthemoreconservativemutationselectioncriteria(termedhereNewtree).


Primary

Lung2.R1

Lung1.R1

Liver2.R2

Liver1.R2

Germline

Patient 1: Old tree

Primary

Lung2.R1

Lung1.R1

Liver2.R2

Liver1.R2

Germline

Patient 1: New tree


Colon2.R1

Primary

ALN

Colon1.R1

Germline

Patient 2: Old tree

Colon1.R1

Primary

ALN

Colon2.R1

Germline

Patient 2: New tree


ALN

Primary

Bone.R1

Germline

Patient 3: Old tree

ALN

Primary

Bone.R1

Germline

Patient 3: New tree


Colon.R4Primary1Primary2Primary3Primary4Uterus.R2Primary5Primary6Brain.R3Germline

Patient 4: Old tree

Uterus.R2Primary1Primary5Primary2Primary4Primary3Primary6Colon.R4Brain.R3Germline

Patient 4: New tree


Primary1

Primary2

Bone.R2

Bone.R1

Germline

Patient 5: Old tree

Primary2

Primary1

Bone.R2

Bone.R1

Germline

Patient 5: New tree


Skin1.R1

Skin2.R1


Germline

Patient 7: Old tree

Skin2.R1

Skin1.R1


Germline

Patient 7: New tree


Primary

ALN

Skin

Bone

Germline

Patient 8: Old tree

ALN

Primary

Skin

Bone

Germline

Patient 8: New tree


Primary

Brain1.R1

Brain2.R1

Brain3.R1

Germline

Patient 9: Old tree

Primary

Brain1.R1

Brain2.R1

Brain3.R1

Germline

Patient 9: New tree


Primary1Skin1.R2Skin2.R2Skin3.R2Primary2ALN2ALN1Germline

Patient 10: Old tree

ALN2Primary1Skin2.R2Skin1.R2Skin3.R2Primary2ALN1Germline

Patient 10: New tree


Brain1.R1Primary1Primary2Primary3Primary4Primary5Brain2.R1Germline


Brain1.R1Primary1Primary2Primary4Primary3Primary5Brain2.R1Germline





ALN.relapseBone2.R1

Bone1.R1

Germline




ALN.relapseBone2.R1

Bone1.R1

Germline



ALN

Primary

Brain2.R1

Brain1.R1

Germline


ALN

Primary

Brain2.R1

Brain1.R1

Germline



Primary

ALN

Liver.R1

Germline


Primary

ALN

Liver.R1

Germline



Skin1.R1Skin2.R1Skin3.R1Primary.local.recurrence.3Primary.local.recurrence.2Primary.local.recurrence.1Germline


Skin3.R1Skin2.R1Skin1.R1Primary.local.recurrence.2Primary.local.recurrence.3Primary.local.recurrence.1Germline



Primary1

Primary2

ALN2

ALN1

Brain.R1

Germline


Primary1

Primary2

ALN2

ALN1

Brain.R1

Germline



ALN1Skin2.R2Skin1.R2Primary1Primary2ALN2Germline


Skin2.R2Skin1.R2ALN1Primary2Primary1ALN2Germline



Primary

Brain2.R5

Brain1.R5

Brain1.R3

Germline


Primary

Brain2.R5

Brain1.R5

Brain1.R3

Germline



Ovary1.R2Ovary2.R2Ovary3.R2Ovary4.R2Ovary5.R2Ovary7.R2Ovary6.R2Primary1Primary2Primary3Primary4Ovary9.R2Ovary8.R2Germline


Ovary1.R2Primary2Primary1Primary3Primary4Ovary3.R2Ovary7.R2Ovary2.R2Ovary4.R2Ovary6.R2Ovary5.R2Ovary9.R2Ovary8.R2Germline



Efficient filtering of FFPE-related C>T/G>A artifacts

To account for potential artifacts induced by formalin-fixed paraffin embedded (FFPE) samples, we employed mutation-

filtering criteria described in the “Variant calling, filtering, and copy number alteration detection” subsection in Methods.

Apartfromotheranalysis,thisalsoinsuredthattheageofFFPEsamplescouldnotnegativelyinfluencethesignatureanalysis.

Toshowhoweffectiveourfilteringwas,wedividedoursamplesintotwogroups,i.e.<=2004and>2004.Thenwecompared

the number of C>T/G>A substitutions between the two groups before and after the filtering. According to the results, the

significantdifferenceinnumberofC>T/G>Asubstitutionsseenbeforefilteringwasefficientlycanceledafterfiltering.

ThefigureshowsNumberofC>Tmutationsbeforeandafterfiltering.Thep-valuesreportedaretwo-sided.


50

100

150

200

<= 2004 (n = 50) > 2004 (n = 49)Year of fixation

Num

ber o

f C>T

mut

atio

ns b

efor

e fil

terin

g p−value = 0.03

40

80

120

160

<= 2004 (n = 50) > 2004 (n = 49)Year of fixation

Num

ber o

f C>T

mut

atio

ns a

fter f

ilter

ing

p−value = 0.86Supplementary Figure 12

0.4

0.6

0.8

Chemotherapy (n =10) No chemotherapy (n = 5)

Frac

tion

of m

etas

tasi

s−sp

ecifi

c m

utat

ions

p−value = 0.51A

0.4

0.6

0.8

Hormonal Therapy (n =9) No Hormonal Therapy (n = 6)

Frac

tion

of m

etas

tasi

s−sp

ecifi

c m

utat

ions

p−value = 0.33B

0.4

0.6

0.8

Radiotherapy (n =12) No Radiotherapy (n = 3)

Frac

tion

of m

etas

tasi

s−sp

ecifi

c m

utat

ions

p−value = 0.45CSupplementary Figure 13

0.2

0.4

0.6

0.8

< 2002 (n = 9) >= 2002 (n = 10)Year of fixation

Frac

tion

of s

ampl

e−sp

ecifi

c m

utat

ions

(lea

ves)

p−value = 0.55A

0.2

0.4

0.6

0.8

Coverage > 79 (n = 9) Coverage < 79 (n = 10)Sequencing Coverage

Frac

tion

of s

ampl

e−sp

ecifi

c m

utat

ions

(lea

ves)

p−value = 0.36B


SUPPLEMENTARY FIGURE LEGENDS

Supplementary Figure 1: Exome sequencing coverage statistics. Stacked bar showing the percentage of target

regions covered at certain coverage. Each bar represents one sample and the bars are grouped by patient.

Supplementary Figure 2: Schematic representation of treatment history, number of relapses, relapse

locations, sequenced relapsed sites, PAM50 intrinsic molecular subtypes for primary tumors and survival

timeline of patients in our cohort. Color bands, whose length is proportional to the timescale, represent different

treatment types. Each molecular subtype is represented by its own specific color. Failed and un-sequenced samples

are colored grey. L+, positive axillary lymph node; Loc, Local relapse; Contr: contralateral event; BL, Basal like; LA,

Luminal A; LB, Luminal B; H2, Her 2 enriched; NBL, Normal breast like CT, chemotherapy; RT, radiotherapy; HT,

hormonal therapy; M1, metastasis 1; M2, metastasis 2; M3, metastasis 3; M4, metastasis 4;

Supplementary Figure 3: Analysis pipeline for investigating tumor progression models in breast cancer.

Given the exome-sequencing data, Mutect was used for calling somatic mutations while AscatNGS was used for

estimating tumor purity and copy number aberrations. The input to phylogenetic reconstruction, using Dollo

Parsimony, consisted of a binary matrix obtained by first weighing the mutant allele frequency by tumor purity and

then thresholding the resulting values by 0.05. To infer the statistical support of internal vertices, non-parametric

bootstrapping was used. The phylogenetic analysis resulted in a tree with bootstrap support. The input to subclonal

reconstruction (using PyClone) consisted of mutant allele frequency, copy number aberrations and tumor purity data.

The subclonal analysis resulted in inferred clusters, represented here as density plot, which shows the cellular

prevalence of each cluster (or subclone) in each sample. Finally the output from phylogenetic and subclonal analysis

is integrated as a tree containing the subclonal information as colored (single clone) or dotted (multiple clones) lines

along its edges. Edge lengths in the tree are scaled by number of substitutions while internal vertices are marked with

bootstrap support values.

Supplementary Figure 4: Separating property in the tumor tree. (A) Inferring the role of axillary lymph node in

seeding distant metastasis, based on the separating property in tumor tree. We observe that Germline-to-Primary

path (color red) is separating the path from “axillary lymph node” to “Metastasis 1”, “Metastasis 2a”, and “Metastasis

2b”. Thus, we infer that Primary, rather than Lymph node, has seeded distant metastases. (b) Inferring linear

progression based on the separating property in tumor tree. We observe that Germline-to-Primary path (colored red)

is not separating the path from “Metastasis 1” to the two blocks of “Metastasis 2”, namely “Metastasis 2a” and

“Metastasis 2b”. Thus we infer that “Metastasis 1”, rather than Primary, has seeded “Metastasis 2”.

Supplementary Figure 5: Pairwise mutation heatmaps for each patient in the cohort. For each patient, the

fraction of shared and specific mutations is presented in the left and right column respectively. The heatmap in the

left column illustrates, for row i and column j, the fraction of shared mutations between i and j divided by the total

mutations in both samples. The heatmap on the right illustrates, for row i and column j, the fraction of specific

mutations present in sample i but absent in present j. Pairwise mutation heatmaps are not given for patient 6 and 12

due to low number of samples. (A) Patient 1 (B) Patient 2 (C) Patient 3 (D) Patient 4 (E) Patient 5 (F) Patient 7 (G)

Patient 8 (H) Patient 9 (I) Patient 10 (J) Patient 11 (K) Patient 13 (L) Patient 14 (M) Patient 15 (N) Patient 16 (O)

Patient 17 (P) Patient 18 (Q) Patient 19 (R) Patient 20.

Supplementary Figure 6: Phylogenetic trees and subclonal information for each patient in the cohort. For

each patient, subclonal information is embedded in the phylogenetic, which is presented as subfigure I. In the tree,

edge lengths are proportional to the number of mutations, with the actual number given in parenthesis for each edge.

The list of alterations in putative driver genes are given for each edge, with mutations, amplifications and deletions

shown in black, red and blue color respectively. The information about individual subclones is given in tabular format

which includes cluster ID of the subclone, the color used for subclone in phylogenetic tree (and the density plot),

number of mutations in the subclone and list of putative driver genes included in the subclone. The density plot is

given as subfigure II, which shows the cellular prevalence of each subclone in each sample. In the density plot, the

cluster IDs along with the number of mutations are given on x-axis while their cellular prevalence in samples are

given on y-axis. Figures for the patients are given in numeric order and exclude patients 6 and 12 since the number

of samples, in both cases, is less than 3 (minimum number for reconstructing a phylogenetic tree). (A) Patient 1 (B)

Patient 2 (C) Patient 3 (D) Patient 4 (E) Patient 5 (F) Patient 7 (G) Patient 8 (H) Patient 9 (I) Patient 10 (J) Patient 11

(K) Patient 13 (L) Patient 14 (M) Patient 15 (N) Patient 16 (O) Patient 17 (P) Patient 18 (Q) Patient 19 (R) Patient 20.

Supplementary Figure 7: (A) i. Residuals sum of squares (RSS) as a function of number of signatures attempted.

Dots represent mean values and bars represent standard errors. ii. Explained variance as a function of number of

signatures attempted. Dots represent mean values and bars represent standard errors. (B) A heatmap of Euclidean

distances between the extracted four signatures (x-axis) and the published signatures (y-axis). (C) Barplots showing

the frequencies of six classes of substitutions in both the transcribed strand (red) and the untranscribed strand (blue)

across the four extracted signatures.

Supplementary Figure 8: Heatmap showing the copy number landscape across samples for each patient. To

visualize the varying landscape of copy numbers between different samples, copy number heatmap for each patient

is given.

Supplementary Figure 9: Lineage analysis for all patients performed using LICHeE. In order to validate our

tumor progression results, we used LICHeE to reconstruct lineage trees for all patients (except patient 6, which has

only a single sample) as described in “Validation of phylogenetic trees” subsection in Methods. LICHeE uses variant

allele frequencies of somatic mutations to reconstruct multi-sample cell lineage trees and infer the subclonal

composition of the samples. (A) Patient 1 (B) Patient 2 (C) Patient 3 (D) Patient 4 (E) Patient 5 (F) Patient 7 (G)

Patient 8 (H) Patient 9 (I) Patient 10 (J) Patient 11 (K) Patient 12 (L) Patient 13 (M) Patient 14 (N) Patient 15 (O)

Patient 16 (P) Patient 17 (Q) Patient 18 (R) Patient 19 (S) Patient 20.

Supplementary Figure 10: Mutations heatmap for all patients to visualize the shared and specific mutations

among samples

Supplementary Figure 11: Comparison of phylogenetic trees reconstructed using two different mutation-

filtering criteria. Validation of phylogenetic trees by removing mutations affected by variable coverage and/or

different tumor purity among samples as described in “Validation of phylogenetic trees” subsection of the Method

section. For each patient, a side-by-side comparison of the tree reported in the paper (termed here Old tree) vs. the

one reconstructed using the more conservative mutation selection criteria (termed here New tree) is presented.

Supplementary Figure 12: Efficient filtering of FFPE-related C>T/G>A artifacts. Boxplots showing the number

of C>T/G>A mutations (Y-axis) in two groups of samples (x-axis) defined based on sample age. The comparison was

performed both before applying filtering (A) and after applying filtering (B). P-values are computed based on two-

sided Mann-Whitney test.

Supplementary Figure 13: Effect of treatment on fraction of metastasis-specific mutations. Boxplots showing

the proportions of metastasis-specific mutations (Y-axis) between two groups of patients divided based on treatment

history. Each type of treatment was tested separately: (A) chemotherapy, (B) hormonal therapy and (C)

radiotherapy. Only patients where at least one primary and one distant metastasis sample have been sequenced (15

patients) were considered for this comparison. For each patient, the fraction of metastasis mutations that are not

detected in primary tumor was computed for each distant metastasis site. In patients where more than one distant

metastasis site were sequenced, we chose the highest fraction. P-values were computed using two sided Mann-

Whitney test.

Supplementary Figure 14: Effect of sample age and coverage on fraction of sample-specific mutations

(length of leaves in phylogenetic trees). Boxplots showing the proportion of sample-specific mutations (Y-axis) for

two groups of patients divided based on time of first sample acquisition (A) and based on average sample coverage

(B). P-values are computed based on two-sided Mann-Whitney test.

Documents

SUPPLEMENTARY METHODS Patient material selection criteria