www.sciencemag.org/cgi/content/full/science.

aaf2288/DC1

Supplementary Materials for

Targeting of cancer neoantigens with donor-derived T cell receptor
repertoires

Erlend Strønen, Mireille Toebes, Sander Kelderman, Marit M. van Buuren,

Weiwen Yang, Nienke van Rooij, Marco Donia, Maxi-Lu Böschen,
Fridtjof Lund-Johansen, Johanna Olweus,* Ton N. Schumacher*

*Corresponding author. E-mail: johanna.olweus@medisin.uio.no (J.O.); t.schumacher@nki.nl (T.N.S.)

Published 19 May 2016 on Science First Release

DOI: 10.1126/science.aaf2288

This PDF file includes:

Materials and Methods
Figs. S1 to S8
Tables S1 to S6
Captions for tables S7 and S8
References

Other supplementary material for this manuscript includes the following:

Tables S7 and S8 (Excel format)
Materials and Methods
Primary cells and cell lines
This study was approved by the Regional Ethics Committee and the Institutional Review Board and
performed in accordance with the Declaration of Helsinki. Tumor tissue and TIL were obtained from
individuals with stage IV melanoma in accordance with national guidelines, where applicable following
signed informed consent and after approval of the local medical ethical committee. HLA-A*02:01pos
healthy donor peripheral blood mononuclear cells (PBMCs) were obtained from buffy coats from the
Sanquin and Oslo University Hospital blood banks. Genomic HLA-typing was performed by the
Norwegian Bone Marrow Donor Registry (WMDA/EFI accredited; Table S6). PBMCs were isolated by
density-gradient centrifugation with Lymphoprep (Axis-Shield). Epstein Barr-virus transformed
lymphoblastoid cell lines (EBV-LCL) were generated from 20 x106 PBMCs through standard protocols
(22). The T2 cell line was obtained from American Type Culture Collection (ATCC). Patient-derived
melanoma lines from patient 1 and patient 3 were previously established from tumor fragments. The
melanoma cell line Mel 526 was obtained from Dr. S. Rosenberg (NIH, Bethesda, USA). FLYRD18
packaging cells have been previously described (ECACC no. 95091902). Patient-derived melanoma
lines were maintained in IMDM (Thermo Fisher) supplemented with 10% FCS and antibiotics, all other
cell lines were maintained in RPMI 1640 supplemented with 10% FCS and antibiotics.

Sequencing and neo-antigen selection

Tumor-specific non-synonymous mutations within expressed genes were identified from three
advanced-stage melanoma patients (Patient 1 (NKIRTIL006, Sanger ID PD9024); Patient 2
(MM131207, Sanger ID PD9031); Patient 3 (MM090904, Sanger ID PD9029). DNA and RNA
sequencing data have been deposited in the European Genome-Phenome Archive (accession codes
EGAD00001000243 and EGAD00001000325).
Exome data of patient 001 were analyzed as follows: Reads were aligned to the human reference
genome GRCh37 using BWA version 0.5.10 (23). Aligned reads were deduplicated using Picard
(http://picard.sourceforge.net/) and realignment around indels was performed using GATK toolkit (24).
Somatic single nucleotide variants (SNV) were called using Somatic-sniper (25) and filtered using a
somatic score cut-off of > 34 and a minimum of 4 reads in both tumor and control. Cut-offs were chosen
based on analysis of exome data in van Rooij et al. (6), showing that 1) A decrease in the stringency of
filtering to a somatic score cut-off of 5 yielded only an additional 20% of mutations. 2) The large
majority of these additional putative mutations was likely to represent noise, as the UV-induced DNA
mutation spectrum that is dominant for mutations with a somatic score cut-off of > 34 is absent for
mutations with a somatic score of 5<m< 34 (only 26% C>T/G>A mutations). To further evaluate the
quality of the selected set of mutations for the exome described within (6), a subset of 78 mutations was
validated using Sanger sequencing. In 72/78 cases the mutation could be confirmed, corresponding to a
false discovery rate of 0.07. The thresholds identified in this manner were also applied to patient 001
described in this study. Somatic indels were called using the GATK somatic indel detector and filtered
using a minimum variant frequency of 25% and having at least 5 reads showing the indel. Germ-line
variants in the vicinity of detected somatic variants were identified using Samtools (26) and filtered
using minimum coverage and minimum number of alternate reads of 10 and 6 reads, respectively.
SNPeff (27) was then used to predict the effect of all variants on the Ensembl gene build version 65.
Using a custom perl script and the Ensembl API, coding variants were edited into the cDNA sequence
and subsequently translated into protein sequence. Protein sequences were separately generated for
normal (only germline variants) and tumor samples (germline variants and tumor specific mutations).

2
RNA was isolated using the TriZol protocol (Invitrogen). Poly-A selected RNA libraries were prepared
using the TruSeq RNA library protocol (Illumina), and the resulting libraries were sequenced on an
Illumina HiSeq2000 using 51bp paired-end reads. Reads were aligned to human reference genome
GRCh37 using Tophat 1.4 (28). Expression values were calculated as FPKM using Cufflinks (29).
Exome and RNA data of patients 002 and 003 were obtained and analyzed as described in Behjati S, et
al. (30). For all three tumors, obtained lists of mutations show a dominant UV signature (78-85% of
mutations, Table S7).

Selection of peptides for TIL screens

Amino acid stretches of 39 aa with the mutated amino acid at position 19 were used to perform
predictions of proteasomal cleavage (NetChop (31)) and MHC class I binding (NetMHC3.2 or
netMHCpan2.0, see Table S8 (32, 33)). For those mutations located within 19 amino acids from the N-
or C-terminus of the protein, shorter fragments were used. For analysis of neo-antigen specific T-cell
reactivity in TIL, potential neo-epitopes (9-11 mers) were identified as described in van Buuren et al.
(34), selecting those epitopes that 1) contain the mutated amino acid; 2) were derived from a gene with
FPKM values > 0; 3) have a predicted C terminal cleavage probability of > 0.5; 4) have a low
‘similarity-to-self’, as described in van Buuren et al. (34). From this set, epitopes were then selected on
the basis of predicted MHC binding affinity, such that per 100 mutations, a total of 35 peptides/allele (a
sum of 9-, 10- and 11-mers) were retained. For HLA-A*02:01, this resulted in selection of peptides with
predicted binding affinity of < 954 nM, < 381 nM and < 3,734 nM for patient 1, 2 and 3, respectively.
Cut-offs in predictions were on purpose set conservatively to minimize the number of false negatives.

Selection of peptides for T- cell stimulations

Peptides for patient 1 (10 nonamers and 10 decamers) and patient 3 (10 nonamers) were selected on
the basis of highest predicted MHC binding affinity for peptides of that length, from genes with RNA
expression value of > 10 FPKM. Peptides for patient 2 (15 nonamers and 12 decamers) were selected on
the basis of highest predicted MHC binding affinity for peptides of that length, from genes with RNA
expression value of > 0 FPKM. Peptides and predicted affinities are listed in Tables S1-S3. Receiver-
operating characteristic (ROC) curves were generated using NetMHC 4.0, representing the most recent
version of this prediction algorithm. All peptides were synthesized by the NKI peptide synthesis facility.

Minigene design
Minigenes for induction of T-cell responses were designed to encode 10-21 potential T-cell
epitopes without flanking sequences. The resulting protein sequences were analyzed with the NetChop
3.1 cleavage prediction algorithm to evaluate whether sites at the transition between all individual
predicted peptides were predicted to be cleaved. In case a certain epitope was not predicted to be
cleaved, epitopes were either shuffled, or K, QLGL or GVGT linker sequences were included up- or
downstream of the epitope to allow cleavage. Minigenes, including Kozak sequence and stop codon,
were codon-optimized and synthesized by Genscript. Subsequently, minigenes were cloned into the
pCIpA102 vector for in vitro mRNA transcription using the RiboMAX LargeScale RNA production
system (Promega), as previously described (16, 35). The minigenes yielded the following products:

Patient 1 nonamers
MK NLNCCSVPV ALSPVIPHI YLVDSVAKM WLIDMKSLV LLDSLPMDV SLFALGNVI
KINLSLFAL FALGNVISA STAFDFLAV ALCGQCVRI

3
Patient 1 decamers
M KLYGLDWAEL FMTKINDLEV ALDPHSGHFV QLGL GLLPVLSWLL ALLETPSLLL
FALGNVISAL HMMGFRTQEV YLAPQESYGA QLGL KLFEFLVYGV QLGL FLVYGVRPGM

Patient 2 nonamers
M MLLAIPEAV FLSQRVAFL FLVFYVIPL SMLHFVTSI ALSPVIPHI GLMVIAWFI FQPSFSHLV
WLPNHVIYL FLMASISSF AVGSYVYSV ILAALFLGL LIIPFIHLI AQFKGAWIL LLFGMPPCL
FLQFRGNEV LMASISSFL GSLDVLMAV FIGPLMDAL KLSHQLVLL GMPPCLLAL
QQFAVGSYV

Patient 2 decamers
M FLMASISSFL FAVGSYVYSV ILPFTPEFLV FVFKLYYFEI LMASISSFLL YVIPLSIISV
SMLLAIPEAV FIAKNLIQSA AILPFTPEFL VTFAVSFYVV GLFRPYGSTV GTAWLEWWHV
KLPVGSQCLV SIISVYYYFI LLGSSCAPPL SLSSSGAPSL LMAVAILKEV

Patient 3 nonamers
M SLDLTTSPV TQFWCAVAL VMDSIDVKI TLAYSFQSL ILDVCSSGL KVDIFFLGL
VMKFKNPLV VTYSGKFLI ALAQKGVQL GVGT ALALAQKGV QLGL

Control minigene
MK ALKDVEERV LLFGLALIEV GLYDGMEHL SLFLGILSV SLLMWITQA MLMAQEALAFL
MLAVISCAV

Induction of neo-antigen reactive T cells and generation of CTL clones and lines
Generation of mature monocyte-derived dendritic cells (MoDCs) and induction of antigen-specific
T cells was carried out as previously described for the induction of allo-reactive T-cell responses (16,
35), but with minor changes. Briefly, mature MoDCs from HLA-A*02:01pos donors were transfected
with 30-60 µg minigene mRNA prior to co-culture with the non-adherent fraction of the autologous
PBMCs at a ratio of 0.6x106 MoDCs to 6x106 PBMCs in 2 ml medium per well in 24-well plates in
CellGro DC medium (CellGenix) supplemented with 10 ng/ml interleukin 7 (PeproTech) and 50 pg/ml
interleukin 12 (PeproTech). 50% of medium was changed on day 7, and on day 12 of co-culture cells
were restimulated with HLA-A*02:01pos EBV-LCL transfected with 20-40 µg minigene mRNA, at a
density of 0.6x106 mRNA-transfected EBV-LCL to 3x106 T cells in 2 ml of medium per well in 24-well
plates and cultured in X-Vivo 20 (Lonza) supplemented with 5% human pooled serum (HS; Trina
biotech) and 10 IU/ml interleukin 2 (IL-2; R&D Systems). On day 19, cultures were analysed for the
presence of pMHC multimerpos CD8 cells restimulated, as described for day 12. On day 26, pMHC-
multimerpos CD8 cells were sorted. To generate neo-antigen reactive CTL clones, 96-well tissue-culture
treated plates were prepared with irradiated allogeneic feeder cells (1-2x105 per well) in 200 µl X-Vivo
20 medium supplemented with 5% HS, 1 µg/ml phytohaemagglutinin (PHA; Remel Thermo Scientific),
100 IU/ml IL-2 and 2 ng/ml interleukin 15 (IL-15; Peprotech), and pMHC-multimerpos CD8 cells were
sorted on a FACS Aria II cell sorter (BD Biosciences) as single-cells into the 96-well plates. To generate
CTL lines, pMHC-multimerpos CD8 cells were sorted in bulk and expanded with irradiated allogeneic
feeder cells at a ratio of 1:200 in X-Vivo 20 supplemented with 100 IU/ml IL-2, 2 ng/ml IL-15, and 30
ng/ml anti-CD3 (OKT-3) or 1 µg/ml PHA. To prepare allogeneic feeder cells, PBMCs from three donors
were irradiated with 30 Gy, washed once and pooled at a 1:1:1 ratio. For antigen-independent expansion

4
of CTL clones and lines, medium was refreshed as needed and at least every 5-6 days. Presence of
relevant cells was confirmed by staining individual CTL clones and lines with pMHC-multimers labeled
with two fluorochromes, and only cultures with double pMHC-multimerpos and CD8pos cells were used
for further analysis.

Fluorescent pMHC multimers, antibodies and flow cytometry

pMHC multimers labeled with phycoerythrin (PE), allophycocyanin (APC), Quantum dot 655 or
BrilliantViolet 421 were prepared and used as described previously (36, 37). To increase specificity of
staining, each pMHC multimer was labeled with two fluorochromes, and only live CD8 cells staining
positively for relevant pMHC-multimer (positive for both fluorochromes), but not for irrelevant pMHC-
multimers, were considered positive (37). In general, more than 1x105 CD8 cells were analyzed, and
cut-offs were set to >0.005% of CD8 cells and more than ten events. Cells or microspheres were labeled
with the following antibodies, in-house conjugated to Alexa Fluor 647, Alexa Fluor 700 or Pacific Blue
(all Life Technologies): anti-CD107a (H4A3), -CD107b (H4B4) (both BD Biosciences), -CD8 (Hit8a;
Tonbo Biosciences), and -HLA-A2 (BB7.2; AbD Serotec), or with -interferon–γ (IFN-γ) PE
(eBioscience), -mouse TCRβ chain PE (H57-597), -IFN-γ PE, -IFN-γ APC, -CD4 FITC, -CD4 PE, -
CD137 PE and -CD107a APC (all BD Biosciences), -CD8 Alexa Fluor 700 (Life Technologies), or -β2-
microglobulin PE (B2M-01; Exbio). Live/Dead Fixable Near-IR or LIVE/DEAD Fixable Violet Dead
Cell Stain kits (Life Technologies) were used for dead-cell exclusion. Samples were analyzed on FACS
Calibur, LSRII or FACS Canto II flow cytometers (BD Biosciences) and data analyzed using FlowJo
(TreeStar) or FACS DIVA (BD Biosciences) software.

Functional analysis of T-cell responses

Functionality of neo-antigen specific T cells was determined by flow cytometric analysis of
CD107a/b mobilization (degranulation) and/or IFN-γ production, as previously described (16, 35), with
the following modifications: to determine peptide sensitivity of CTL clones, clones were bar-coded by
labeling each clone with a unique combination of Carboxyfluorescein, succinimidyl ester (CFSE) and
CellTrace Violet (both Life Technologies) concentrations, before pooling of up to 16 clones into one
sample and subsequent incubation with target cells. T2 cells were pulsed with the indicated peptide
concentrations for one hour at 37oC, washed, and then incubated for five hours with bar-coded CTL
clones. To assess tumor cell recognition, CTL clones were incubated for five hours with patient-derived
melanoma cell lines expressing either the relevant mutated gene, or the WT counterpart. Where
indicated, tumor cell lines were pulsed with 1 nM of the relevant neo-antigen for one hour prior to
incubation. In all assays, background degranulation of CTL clones in medium without stimulator cells
was measured and subtracted from all samples. When CD137 was used as an activation marker, target
cells were transfected with mRNA or pulsed with peptide as indicated, followed by incubation with CTL
lines for 24 hours. Following incubation, cells were harvested and stained with anti-CD8 BrilliantViolet
421, anti-CD137 PE and live/dead cell marker, prior to analysis.

Amplification, sequencing and cloning of TCR genes

RNA was isolated from 28 T-cell clones and processed individually according to a modified single-
cell based PCR protocol by Tang et al. (38) to obtain TCR-specific cDNA. In brief, an RT-PCR was
performed for each sample using four pairs of TCRα/β constant-domain specific primers after which
double-stranded DNA was obtained by adding a poly-G tail to the first strand and, following RNA
degradation, a template switch to synthesize the second strand. Finally, two rounds of nested PCR
amplification were performed using additional constant domain primers and adaptor primers annealing

5
to an anchor sequence introduced in the poly-G domain. Libraries were made using the Kappa Illumina
kit and subsequently sequenced on an Illumina MiSeq. Sequence data were analysed using the MiTCR
script that identifies the CDR3 region between conserved serine and phenylalanine residues on
chromosome 7 and 14 (39). Full-length TCR chains were reconstructed using an in-house developed
python script, TCRprimer, which reads upstream from the CDR3 region into the variable domain and
generates a consensus sequence by cross-linking the input sequence data to a database containing all
human TCR variable domains. Output was manually verified for each sample in IMGT/V-Quest (40). A
variant pMP71 retroviral backbone for TCR cloning was utilized to rapidly exchange TCR V domains
(39). To maximize TCR expression, this vector contains codon-optimized mouse TCRαβ constant
domains with additional Cys residues and a porcine teschovirus–derived P2A sequence to link TCR
chains. Variable TCRα and TCRβ fragments of identified TCRs were codon optimized, synthesized and
cloned by Genscript.

Retroviral transduction of T-cell receptors

For retrovirus production, FLYRD18 packaging cells were plated into 6 well plate dishes at
0.5×106 cells per well. After 24 hours, cells were transfected with 3 µg retroviral vector DNA, using 9 µl
of X-tremeGENE 9 Transfection reagent (Roche Diagnostics). PBMCs were incubated for 30 minutes
with Cell Therapy Systems Dynabeads CD3/CD28 (Thermo Fisher Scientific Inc.), at a bead:CD3pos cell
ratio of 1:1. CD3/CD28pos cells were magnetically enriched and plated at 1.5 × 106 cells per well in 24-
well plates with 100 IU/ml rh-IL-2 and 5 ng/ml rh-IL-15 (Peprotech). After 48 hours, 0.2 - 0.3 × 106
cells were resuspended in 1 ml retroviral supernatant plus 1 ml medium, supplemented with rh-IL-2 (100
IU/ml final) and rh-IL-15 (5 ng/ml final), and transferred to Retronectin (Takara)-coated plates. Plates
were centrifuged for 90 minutes at 430g. Transduction efficiency was determined at 72 hours by staining
with pMHC multimer and anti-mouse TCRβ chain antibody. Two weeks after transduction, cells were
tested for antigen reactivity.

Introduction of mutant antigens into third party melanoma cells

pMX-IRES-GFP retroviral vectors carrying mutant antigen-encoding minigenes were generated
such that predicted neo-antigens were flanked by 15 amino acids of the natural sequence on either side
of the predicted antigen, yielding the following translation products:
ASTN1P>L: MILRRSSLKYLGCRYSEIKLYGLDWAELSRDLRKTCEEQTLSIPYN
CDK4R>L: MRYEPVAEIGVGAYGTVYKALDPHSGHFVALKSVRVPNGGGGGGGLP
SMARCD3H>Y: MAADEVAGGARKATKSKLFEFLVYGVRPGMPSGARMPHQGAPMG
MLL2L>H: MEAPRFPHLGSGRWEQEDRALSPVIPHIPRASIPVFPDTKPYGALG.
Sequence in bold indicates the predicted neo-antigen, underlined residue indicates the mutant
amino acid. To create modified melanoma lines, FLYRD18 packaging cells were plated into 6 well plate
dishes at 0.5×106 cells per well. After 24hrs, cells were transfected with 3 µg retroviral vector DNA
using 9 µl of X-tremeGENE 9. Virus supernatant was harvested and added to the melanoma lines in the
presence of 8 ng/ml polybrene and cells were incubated for 72 hours. Transduction efficiency was
measured by flow cytometry, and GFP positive cells were sorted on a MoFlo cell sorter (Beckman
Coulter, Inc.) to more than 90% purity.

6
CRISPR/Cas9 knock out of mutant MLL2 gene
Single-guide RNA (sgRNA)-encoding oligonucleotides were individually designed (http://tools.genome-
engineering.org) and synthesized (Integrated DNA Technologies) to allow targeting of the MLL2 gene.
CACCG AGCCCAGATGAGGGAAACGA MLL2#1 for
AAAC TCGTTTCCCTCATCTGGGCT C MLL2#1 rev
CACCG ATGCTCTCAGGGGATGAAGC MLL2#2 for
AAAC GCTTCATCCCCTGAGAGCAT C MLL2#2 rev
Guide RNA-encoding sequences were cloned into pSpCas9n(BB)-2A-GFP (Addgene plasmid ID:
48140), as previously described (41), and resulting constructs were validated by Sanger sequencing.
Tumor cells (patient 1) were transfected with both guides using X-tremeGENE, and GFP positive cells
were sorted at 200 cells/ well, a concentration at which only some wells yield growing cells. After
expansion, tumor cell lines were analyzed by extraction of genomic DNA (QIAGEN DNeasy
purification kit), and PCR with MLL2 forward and reverse primers. Resulting products from 2 cell lines
were cloned into TOPO TA 2.1 vector and sequenced. Eight independent sequences from one of these
tumor cell lines revealed a frame shift within the mutant MLL2 allele that disrupted the ORF of the
MLL2 neo-antigen.

pMHC stability assay

Soluble, biotinylated HLA-A*02:01 monomers complexed with a conditional ligand sensitive to
UV-light were prepared and UV-mediated peptide-exchange reactions were carried out in 96-well poly-
propylene plates, as described previously (36). Microspheres for MHC immobilization were prepared as
described previously (42) except for the use of streptavidin (Peprotech) rather than protein G. Following
UV-irradiation, reactions were left at 24 hours to reach equilibrium, before peptide-exchanged MHC
complexes were coupled to streptavidin-coated microspheres at a ratio of 16.8 x 106 pMHC molecules
per microsphere, at a pMHC concentration of 250 nM, by incubation at room temperature for 30
minutes. Beads were then washed three times in phosphate-buffered saline supplemented with 1%
Tween and 0.05% bovine γ-globulin (Sigma-Aldrich) to remove excess MHC molecules and peptide,
and either analyzed immediately (time point t=0), or incubated at 37 degrees and analyzed at indicated
time points. For analysis, MHC-coated beads were stained with a conformation dependent anti-HLA-A2
antibody or with anti-β2-microglobulin, and analyzed by flow cytometry to measure unfolding of the
MHC class I heavy chain/dissociation of β2m as a decrease in mean fluorescence intensity. To validate
the assay, a number of well-characterized T-cell epitopes known to bind HLA-A*02:01 were used as
positive controls (Table S5). MHC molecules exposed to UV in the absence of rescue peptide, or MHC
exchange reactions carried out in the presence of an HLA-A*03:01-binding peptide from influenza A
nucleoprotein (Flu NP265-273), were used as negative controls (Table S5). Comparison of data sets
showed that pMHC stability could be measured equally well using anti-HLA-A*02:01 or anti-β2-
microglobulin fluorescence (Fig. S8B). Decrease in anti-β2-microglobulin fluorescence was used in
subsequent experiments, as this assay can also be utilized for HLA class I alleles for which
conformation specific antibodies are not available.

Statistical analysis
GraphPad Prism 6 (GraphPad Software, Inc.) was used to calculate p-values using the Mann-
Whitney U test, to calculate Spearman correlation coefficients, and to create ROC curves where
indicated. ROC curves were compared using the package pROC for R, as previously described (43).

7
 A Irrelevant CT/CD20 Relevant
minigene mRNA minigene mRNA peptide

MAGE A10-
reactive
CTL clone
2% 92% 95%

MAGE C2- CD8 Pacific Blue

reactive
CTL clone
1% 57% 95%

CD107a/b Alexa 647

B Full-length CT/CD20 CD20188-196

CD20 mRNA minigene mRNA peptide
CD8 Pacific Blue

CD20-
reactive
CTL line
87% 92% 91%

CD107a/b Alexa 647

Fig. S1.
Efficient presentation of minigene-encoded T-cell epitopes. A. Mage A10- and Mage C2-reactive
CTL clones were incubated with target cells transfected with relevant (CT/CD20) or irrelevant minigene
mRNA, or with target cells pulsed with 10 µM relevant peptide, followed by measurement of
degranulation responses, shown as CD107 a/bpos events. B. A CD20-reactive CTL line was incubated
with target cells transfected with mRNA encoding the full-length CD20 gene, with relevant (CT/CD20)
minigene mRNA, or with target cells pulsed with 10 µM relevant peptide. All plots are gated on live
CD8 cells.

8
 A B

Mutant
peptide

WT
peptide

Fig. S2
Strategy for multiplexed analysis of CTL clones. A. Data depict SMARCD3H>Y-reactive clones of
donor 2, individually bar-coded with unique combinations of CFSE and CellTrace Violet concentrations,
combined and analyzed for recognition of peptide-pulsed T2 cells by degranulation (CD107a/b). Lower
plots depict data for clones 1-8. B. Representative analysis of degranulation for one CTL clone from a
set of 16 bar-coded MLL2L>H reactive clones following incubation with target cells pulsed with
indicated concentrations of mutant or corresponding WT peptide.

9
 A
No peptide WT peptide Mutant peptide

%CD107a/b+ cells
100 10 pM 1 nM 100 nM 10 pM 1 nM 100 nM
80 CDK4R>L-
60 reactive
40 CTL clones
20
0
1 3 5 7 2 4 6 1 3 5 7 2 4 6 1 3 5 7 2 4 6 1 3 5 7
CTL clone
No peptide WT peptide Mutant peptide
%CD107a/b+ cells

100 10 pM 1 nM 100 nM 10 pM 1 nM 100 nM

80 ASTN1P>L-
60 reactive
40 CTL clones
20
0
1 3 5 7 9 111315 1 3 5 7 9 111315 1 3 5 7 9 111315 1 3 5 7 9 111315 1 3 5 7 9 111315 1 3 5 7 9 111315 1 3 5 7 9 111315
CTL clone
No peptide WT peptide Mutant peptide
%CD107a/b+ cells

100 10 pM 1 nM 100 nM 10 pM 1 nM 100 nM

80 SMARCD3H>Y-
60 reactive
40 CTL clones
20
0
1 3 5 7 9 111315 1 3 5 7 9 111315 1 3 5 7 9 111315 1 3 5 7 9 111315 1 3 5 7 9 111315 1 3 5 7 9 111315 1 3 5 7 9 111315
CTL clone
No peptide WT peptide Mutant peptide
%CD107a/b+ cells

100 10 pM 1 nM 100 nM 10 pM 1 nM 100 nM

80 GNL3LR>C-
60 reactive
40 CTL clones
20
0
1 3 5 7 9 111315 1 3 5 7 9 111315 1 3 5 7 9 111315 1 3 5 7 9 111315 1 3 5 7 9 111315 1 3 5 7 9 111315 1 3 5 7 9 111315
CTL clone

B
No peptide Mutant peptide
%CD107a/b+ cells

100 100 pM 10 nM 1 µM
80 ASTN1 P>L -
60 reactive
40 CTL clones
20
0
1 2 3 4 5 6 7 8 9 1011 1 2 3 4 5 6 7 8 9 1011 1 2 3 4 5 6 7 8 9 1011 1 2 3 4 5 6 7 8 9 1011
CTL clone
No peptide Mutant peptide
%CD107a/b+ cells

100 100 pM 10 nM 1 µM
80 SMARCD3H>Y-
60 reactive
40 CTL clones
20
0
1 5 9 13 17 21 25 29 33 37 41 1 5 9 13 17 21 25 29 33 37 41 1 5 9 13 17 21 25 29 33 37 41 1 5 9 13 17 21 25 29 33 37 41
CTL clone

C
No peptide WT peptide Mutant peptide
%CD107a/b+ cells

100 10 pM 1 nM 100 nM 10 pM 1 nM 100 nM

80 MLL2L>H-
60 reactive
40 CTL clones
20
0
1 5 9 13 17 21 25 29 1 5 9 13 17 21 25 29 1 5 9 13 17 21 25 29 1 5 9 13 17 21 25 29 1 5 9 13 17 21 25 29 1 5 9 13 17 21 25 29 1 5 9 13 17 21 25 29
CTL clone
No pep. WT peptide Mutant peptide
%CD107a/b+ cells

100 10 pM 1 nM 100 nM 10 pM 1 nM 100 nM

80 CDK4 R>L-
60 reactive
40 CTL clones
20
0
1 2 1 2 1 2 1 2 1 2 1 2 1 2
CTL clone
No peptide WT peptide Mutant peptide
%CD107a/b+ cells

100 10 pM 1 nM 100 nM 10 pM 1 nM 100 nM

80 SMARCD3H>Y-
60 reactive
40 CTL clones
20
0
1 6 11 16 21 1 6 11 16 21 1 6 11 16 21 1 6 11 16 21 1 6 11 16 21 1 6 11 16 21 1 6 11 16 21
CTL clone

Fig. S3
Reactivity of donor-derived pMHC-multimerpos CTL with mutant antigen. pMHC-multimerpos
CTL clones were bar-coded with CFSE and CellTrace Violet before incubation with peptide-pulsed

10
target cells. A. The neo-antigen reactive CTL clones (donor 4) shown in Fig. 1 were incubated with
target cells pulsed with corresponding WT peptide or with non-pulsed target cells, as indicated.
Reactivity against cognate neo-antigen (Mutant peptide), as also depicted in Fig. 1, is included for
comparison. Culture conditions and response measurements are identical to those described in the
legend of Fig. 1. B. ASTN1P>L- and SMARCD3H>Y-reactive CTL clones from donor 2 were incubated
with target cells pulsed with the indicated concentrations of the relevant neo-antigen or with non-pulsed
target cells. C. MLL2L>H-, CDK4R>L- and SMARCD3H>Y-reactive CTL clones from donor 3 were
incubated with target cells pulsed with the relevant neo-antigen, with the corresponding WT peptide, or
with non-pulsed target cells. The percentage of CD107a/bpos cells of total CD8 cells is depicted. Values
reflect means of triplicates, error bars indicate SD.

11
 ASTN1P>L-reactive CTL clones SMARCD3H>Y-reactive CTL clones

ASTN1WT tumor ASTN1MUT tumor SMARCD3WT tumor SMARCD3MUT tumor

50 40

% positive cells

% positive cells
40 Donor 2 30 Donor 2
30
20
20
10 10
0 0
1 2 3 4 5 6 7 11 1 2 3 4 5 6 7 11 1 11 13 34 1 11 13 34

ASTN1WT tumor ASTN1MUT tumor SMARCD3WT tumor SMARCD3MUT tumor

40 50
% positive cells

% positive cells
30 Donor 4 40 Donor 4
30
20
20
10 10
0 0
1 3 5 7 9 11 13 15 1 3 5 7 9 11 13 15 1 3 5 6 8 9 10121316 1 3 5 6 8 9 10121316

GNL3LR>C-reactive CTL clones

GNL3LWT tumor GNL3LMUT tumor

100 No peptide 1 nM mut peptide No peptide 1 nM mut peptide

% positive cells

80
60 Donor 4
40
20
0
1 3 5 7 9 11 13 15 1 3 5 7 9 11 13 15 1 3 5 7 9 11 13 15 1 3 5 7 9 11 13 15
CD107a/b+ IFN-γ- CD107a/b+ IFN-γ+ CD107a/b- IFN-γ+

Fig. S4
Tumor recognition by donor-derived ASTN1P>L-, SMARCD3H>Y- and GNL3LR>C-reactive CTL
clones. Clones were incubated with a melanoma line derived from patient 1 that carried the indicated
mutated (MUT) gene, or with a third party HLA-A*02:01pos melanoma line carrying the corresponding
WT gene, and analyzed as shown in Fig. 2. Data shown represent the percentage of IFN-γpos and/or
CD107a/bpos cells among live CD8 cells. Arrows indicate ASTN1P>L- and SMARCD3H>Y-reactive CTL
clones that were selected for TCR sequencing. GNL3LR>C-reactive CTL clones did not respond to
matched or non-matched melanoma cells unless tumor cells were pulsed with the corresponding mutant
peptide. Values reflect means of duplicates, error bars indicate SD.

12
 A B
Relevant mRNA Irrelevant mRNA
0.218%
10
USP28C>F PGM5 H>Y-465-473
SNX24 P>L
MRM1 T>P
Multimer BV421 USP28 C>F
SLC38A1 W>L
1.26%
1 PGM5 H>Y-462-470

SLC38A1W>L

% multimer+ cells of CD8+ cells

Multimer Qd655
0.1
4.73%

PGM5H>Y-37

Multimer BV421 0.01

0.108%
Multimer PE

PGM5H>Y-38

0.001
Multimer Qd655 5 6 7 8
Donor

C USP28C>F- SNX24P>L- PGM5H>Y-462-470- PGM5H>Y-465-473- PGM5H>Y-465-473-

reactive CTLs reactive CTLs reactive CTLs reactive CTLs reactive CTLs
Donor 8 Donor 5 Donor 8 Donor 8 Donor 7
99% pMHC- 80% pMHC- 80% pMHC- 33% pMHC- 88% pMHC-
multimer+ cells multimer+ cells multimer+ cells multimer+ cells multimer+ cells

100 80 80 100
40
% repsponding cells

80 80
of CD8+ cells

60 60 30
60 60
40 40 20
40 40

20 20 20 10 20

0 0 0 0 0
100 1 0.1 100 1 0.1 100 1 0.1 100 1 0.1 100 1 0.1
[nM] [nM] [nM] [nM] [nM]
Mutant peptide WT peptide

Mutated mRNA

WT mRNA

No mRNA/peptide

100 nM peptide

Minigene mRNA

0 20 40 60 80100 0 20 40 60 80 0 20 40 60 80100 0 10 20 30 40 0 20 40 60 80100

% CD137+ cells of CD8+ cells

Fig. S5
Donor-derived neo-antigen specific T-cell responses for patient 2. A. T-cell responses from donor 8
to four predicted neo-antigens from patient 2. Dot plots are gated on live CD8pos cells, colored double
positive populations represent CD8pos cells staining with 2-fluorochrome coded relevant pMHC-
multimers. Values reflect the percentage pMHC-multimerpos cells of total CD8 cells. B. T-cell responses

13
from four healthy donors (donor 5-8) to predicted neo-antigens from Patient 2 after stimulation with
minigene mRNA-transfected APCs. Data depict the percentage of pMHC-multimerpos cells of total CD8
cells. C. CTL lines reactive with USP28C>F (donor 8), SNX24P>L (donor 5), PGM5H>Y-462-470 (donor 8)
and PGM5H>Y-465-473 (donors 7 and 8) were incubated with target cells pulsed with indicated
concentrations of the relevant mutant peptide or the corresponding WT peptide (top row). Values reflect
the percentage of IFN-γpos and/or CD107a/bpos cells of total CD8 cells. CTL lines were also incubated
for 24 hours with target cells transfected with either a minigene construct encoding the predicted neo-
antigens PGM5H>Y, USP28C>F and SNX24P>L, each flanked on both sides by 10 naturally occurring
amino acids (Mutated mRNA), a minigene construct encoding the corresponding WT sequences (WT
mRNA), or the minigene construct used to induce T-cell responses (Minigene mRNA) (bottom row).
Values reflect the percentage of CD137pos cells of total CD8 cells. Percentages of pMHC-multimerpos
cells within each CTL line are indicated. Data represent means of duplicates (top row) or triplicates
(bottom row), error bars indicate SD.

14
 Mock transduced MLL2 TCR 41 CDK4 TCR 53 CDK4 TCR 55 CDK4 TCR 57 CDK4 TCR 17 ASTN1 TCR52 ASTN1 TCR 65 SMARCD3 TCR59 SMARCD3 TCR67

0.6% 52% 28% 35% 15% 37% 44% 11% 56% 31%
Counts

Mouse TCRβ PE

Fig. S6
Transduction efficiency of neo-epitope reactive TCRs. Peripheral blood T cells were transduced with
indicated TCRs, and an aliquot was stained with anti-mouse TCRβ chain antibody to evaluate
transduction efficiency immediately prior to co-culture of T cells with melanoma cells. All histograms
are gated on live CD8pos cells and numbers indicate percentage of CD8pos cells staining positively for
anti-mouse TCRβ. Note that the fraction of TCR-expressing T cells is underestimated by staining with
anti-mouse TCRβ antibody for TCRs 57 and 65 (see Fig. 3B,D).

15
 A TCR17
B TCR55 TCR57
50 TCR53 30 50

% responding cells

% responding cells

% responding cells
% responding cells
30 25
40 25 40
30
20 30
20
15 15
20 10 20
10
10 5 5 10
0 0 0 0
10-6 10-4 10-2 100 102 10-6 10-4 10-2 100 102 10-6 10-4 10-2 100 102 10-6 10-4 10-2 100 102
[nM] [nM] [nM] [nM]

Fig. S7
Peptide-sensitivity of patient- and donor-derived CDK4R>L-reactive TCRs. Healthy donor
peripheral blood T cells transduced with the indicated patient-derived (A) or donor-derived (B) TCRs
were incubated with T2 cells pulsed with the indicated concentrations of CDK4R>L peptide. Graphs
depict means of duplicates, error bars indicate SD.

16
 A No peptide Flu NP265-273 (A3) Flu MP58-66 (A2) gp100280-288 A288V
80000 80000 80000 80000
60000
HLA-A2 MFI
60000

HLA-A2 MFI
60000 60000

HLA-A2 MFI

HLA-A2 MFI
40000 40000 40000 40000
20000 20000 20000 20000
0 0 0 0
0 24 48 72 96 120 144 0 24 48 72 96 120 144 0 24 48 72 96 120 144 0 24 48 72 96 120 144
Time (hours) Time (hours) Time (hours) Time (hours)
MART-126-35 WT MART-126-35 A27L CD20187-195 CMV pp65495-503
80000 80000 80000 80000

60000 60000
HLA-A2 MFI

HLA-A2 MFI
60000 60000

HLA-A2 MFI

HLA-A2 MFI
40000 40000 40000 40000

20000 20000 20000 20000

0 0 0 0
0 24 48 72 96 120 144 0 24 48 72 96 120 144 0 24 48 72 96 120 144 0 24 48 72 96 120 144
Time (hours) Time (hours) Time (hours) Time (hours)

B C
100 1.0
Measured t1/2 β2m (hours)

0.8
Measured t1/2 β2m
Sensitivity

r= 0.9935 0.6 Area under curve: 0.86

10 p<0.0001
0.4 Predicted affinity
Area under curve: 0.72
0.2
p=0.016
1 0.0
1 10 100 0.0 0.2 0.4 0.6 0.8 1.0
Measured t1/2 HLA-A2 (hours) 1 - Specificity

Fig. S8
pMHC stability predicts neo-antigen immunogenicity. A. UV-mediated exchange reactions of
biotinylated soluble, monomeric HLA-A*02:01 molecules were performed in the presence of indicated
HLA-A*02:01-restricted T-cell epitopes, in the presence of a negative control peptide (the HLA-
A*03:01-restricted influenza A NP265 epitope), or in the absence of rescue peptide. Peptides and
predicted affinities are listed in Table S5. Samples were then left for 24 hours at room temperature to
reach equilibrium, combined with streptavidin-coated microspheres and incubated at 37oC for the
indicated time periods. Subsequently, samples were stained with a conformation-dependent anti-HLA-
A2-antibody and analyzed by flow cytometry to measure unfolding of the HLA-A*02:01 heavy chain as
a decrease in fluorescence intensity. Dots represent means of triplicates, error bars indicate SD. B.
Stability of HLA-A*02:01 in complex with the 57 candidate neo-epitopes (Tables S1-S3) and the 6
validated T-cell epitopes (Table S5) was measured either by a decrease in anti-HLA-A2 binding as
described in A, or by a decrease in anti-β2-microglobulin binding, in separate samples. Spearman
correlation between pMHC stability measured by a decrease in anti-β2-microglobulin binding and
pMHC stability measured by a decrease in anti-HLA-A2 binding. D. Receiver-operating characteristic
(ROC) curves illustrating superior predictive value of measured pMHC stability relative to predicted
HLA-A*02:01 binding affinity alone, for the peptides in Tables S1,2,3 and 5. NetMHC 4.0 was used to
predict affinities. Statistical comparison of ROC curves was performed as described in materials and
methods.

17
Table S1. Neo-peptides from patient 1

Wildtype Neo-peptide
Neo-peptide predicted predicted Neo-peptide CD8+ T cell CD8+ T cell
Wildtype amino
Peptide HUGO amino acid binding binding t1/2 β2 response response
acid sequence
ID symbol sequence affinity to affinity to microglobulin observed in induced in
HLA- HLA- (hours) patient donor
A*02:01* A*02:01*
1 GNL3L NLNRCSVPV NLNCCSVPV 49 17 23.5 No Yes
2 MLL2 ALSPVIPLI ALSPVIPHI 13 18 47.7 No Yes
3 MAP2K3 YLVDSVAKT YLVDSVAKM 40 10 17.9 No No
4 AKAP6 WLIDMESLV WLIDMKSLV 14 24 12.3 No No
5 GDAP1 LLDSLPMDA LLDSLPMDV 605 54 6.7 No No
6 KIF3B SLSALGNVI SLFALGNVI 1240 110 3.6 No No
7 KIF3B KINLSLSAL KINLSLFAL 2938 136 4.1 No No
8 KIF3B SALGNVISA FALGNVISA 3352 238 4.6 No No
9 LAMA1 STASDFLAV STAFDFLAV 150 363 3.7 No No
10 SIVA1 ALCGQCVRT ALCGQCVRI 6527 633 5.0 No No
11 ASTN1 KPYGLDWAEL KLYGLDWAEL 16920 43 8.4 No Yes
12 DNAH8 FMTKINGLEV FMTKINDLEV 47 31 15.4 No No
13 CDK4 ARDPHSGHFV ALDPHSGHFV 21304 119 47.5 Yes Yes
14 SLC26A9 GLLPVLSWLP GLLPVLSWLL 678 30 6.4 No No
15 GCN1L1 ALLETLSLLL ALLETPSLLL 31 27 39.9 Yes No
16 KIF3B SALGNVISAL FALGNVISAL 703 160 4.6 No No
17 WDR59 HMMEWFRTQE HMMGFRTQEV 2183 18 7.2 No No
18 PRAMEF11 YPAPQESYGA YLAPQESYGA 19663 39 5.0 No No
19 SMARCD3 KLFEFLVHGV KLFEFLVYGV 10 6 83.7 No Yes
20 SMARCD3 FLVHGVRPGM FLVYGVRPGM 173 102 3.2 No No
*Predicted with NetMHC 3.2.

18
Table S2. Neo-peptides from patient 2

Wildtype Neo-peptide
Neo-peptide predicted predicted Neo-epitope CD8+ T cell CD8+ T cell
Wildtype amino
Peptide HUGO amino acid binding binding t1/2 β2 response response
acid sequence
ID symbol sequence affinity to affinity to microglobulin observed in induced in
HLA- HLA- (hours) patient donor
A*02:01* A*02:01*
21 SSPN FLMASISSS FLMASISSF 15 11.79 4.6 No No
22 FAT4 FVSKLYYFEI FVFKLYYFEI 17 9.57 5.4 No No
23 USP28 LIIPCIHLI LIIPFIHLI 37 28.01 10.0 No Yes
24 SSPN FLMASISSSL FLMASISSFL 3 2.42 17.8 No No
25 SSPN LMASISSSLL LMASISSFLL 19 7.28 6.7 No No
26 ARHGEF12 AISPFTPEFL AILPFTPEFL 101 73.1 4.8 No No
27 TGIF1 LLGSSCAPPP LLGSSCAPPL 5669 30.05 3.2 No No
28 BRWD3 SLSSSGAPSP SLSSSGAPSL 11433 54.63 2.8 No No
29 ENTPD7 FLRQRVAFL FLSQRVAFL 51 5.47 23.5 No No
30 SLC38A1 IWAALFLGL ILAALFLGL 4258 18.6 6.9 No Yes
31 TRPM8 AQSKGAWIL AQFKGAWIL 299 34.06 4.0 No No
32 MRM1 LLFGMTPCL LLFGMPPCL 17 37.68 14.3 No Yes
33 SSPN LMASISSSL LMASISSFL 18 7.28 19.2 No No
34 TARBP1 IIGPLMDAL FIGPLMDAL 1989 95.66 2.9 No No
35 SNX24 KLSHQPVLL KLSHQLVLL 180 42.49 24.2 No Yes
36 MRM1 GMTPCLLAL GMPPCLLAL 21 48.85 5.4 No No
37 ARHGEF12 ISPFTPEFLV ILPFTPEFLV 2097 36.36 3.4 No No
38 TMEM127 VTFAVSFYLV VTFAVSFYVV 26 73.92 4.9 No No
39 MEF2D GLFRRYGSTV GLFRPYGSTV 78 87.28 4.1 No No
40 CSPG4 GTAWLEWRHV GTAWLEWWHV 2153 140.49 4.3 No No
41 TNC KLPVGSQCSV KLPVGSQCLV 86 92.08 4.1 No No
42 GIGYF1 PMAVAILKEV LMAVAILKEV 821 18.74 11.7 No No
43 ZNF169 FQPSFPHLV FQPSFSHLV 16 8.01 10.0 No No
44 AGFG2 FLQSRGNEV FLQFRGNEV 15 23.73 3.8 No No
45 GIGYF1 GSLDVPMAV GSLDVLMAV 351 38.59 5.0 No No
46 PGM5 QQFAVGSHV QQFAVGSYV 441 73.53 5.4 No Yes
47 PGM5 AVGSHVYSV AVGSYVYSV 37 15.89 12.4 No Yes
*Predicted with NetMHCpan 2.0.

19
Table S3. Neo-peptides from patient 3

Wildtype Neo-peptide
Neo-peptide predicted predicted Neo-peptide CD8+ T cell CD8+ T cell
Wildtype amino
Peptide HUGO amino acid binding binding t1/2 β2 response response
acid sequence
ID symbol sequence affinity to affinity to microglobulin observed in induced in
HLA- HLA- (hours) patient donor
A*02:01* A*02:01*

48 GOLGA3 SLDPTTSPV SLDLTTSPV 17 33 13.7 No No

49 PAFAH2 TQFRCAVAL TQFWCAVAL 922 115 4.3 No No
50 GPC4 VMDPIDVKI VMDSIDVKI 140 121 5.1 No No
51 ALS2CR4 TLAYPFQSL TLAYSFQSL 239 126 3.4 No No
52 EPB41 ILGVCSSGL ILDVCSSGL 782 159 3.9 No No
53 EIF2AK3 KVDIFSLGL KVDIFFLGL 509 291 3.9 No No
54 DCI VMKFKNPPV VMKFKNPLV 239 350 3.4 No No
55 HELLS VTNSGKFLI VTYSGKFLI 6662 888 3.7 No No
56 BCS1L ALARKGVQL ALAQKGVQL 1432 914 4.4 No No
57 BCS1L ALALARKGV ALALAQKGV 4246 925 9.2 No No
*Predicted with NetMHC 3.2.

20
 Table S4. T-cell receptor sequences

Donor/
TCR CDR3 amino acid V J CDR3 amino acid V J D
Reactivity Clone
# sequence segment segment sequence segment segment segment
#
TRAV38-2/
ASTN1P>L 2/1 62 CAYFGAQKLVF TRAJ54 CASRPSRGTNYGYTF TRBV6-1 TRBJ1-2 TRBD1
DV8
TRAV38-2/
ASTN1P>L 2/2 63 CAYFGAQKLVF TRAJ54 CASRPSRGTNYGYTF TRBV6-1 TRBJ1-2 TRBD1
DV8
TRAV38-2/
ASTN1P>L 2/3 64 CAYFGAQKLVF TRAJ54 CASRPSRGTNYGYTF TRBV6-1 TRBJ1-2 TRBD1
DV8
TRAV38-2/
ASTN1P>L 2/5 65 CAYFGAQKLVF TRAJ54 CASRPSRGTNYGYTF TRBV6-1 TRBJ1-2 TRBD1
DV8
TRAV38-2/
ASTN1P>L 2/6 66 CAYFGAQKLVF TRAJ54 CASRPSRGTNYGYTF TRBV6-1 TRBJ1-2 TRBD1
DV8
TRAV38-2/
ASTN1P>L 4/2 51 CARYYGQNFVF TRAJ26 CASSERGMVEAFF TRBV25-1 TRBJ1-1 TRBD1
DV8
ASTN1P>L 4/3 52 CALDDRGSTLGRLYF TRAV19 TRAJ18 CASSPRNLGPSGSYEQYF TRBV9 TRBJ2-7 TRBD2

CDK4R>L 3/1 53 CVVSDLYNTDKLIF TRAV8-2 TRAJ34 CASSQNYEQYF TRBV7-8 TRBJ2-7 TRBD2

CDK4R>L 3/2 54 CVVSDLYNTDKLIF TRAV8-2 TRAJ34 CASSQNYEQYF TRBV7-8 TRBJ2-7 TRBD2

TRAV14/ TRBD1,
CDK4R>L 4/1 55 CAMSSLSGNTGKLIF TRAJ37 CASSYSWGAGYEQYF TRBV6-5 TRBJ2-7
DV4 TRBD2
TRBV12-4,
CDK4R>L 4/2 56 CALPYGNKLVF TRAV38-1 TRAJ47 CASTPTGAYEQYF TRBJ2-7 TRBD2
TRBV12-3
CDK4R>L 4/3 57 CAVILRSNDYKLSF TRAV8-1 TRAJ20 CSAGTGELFF TRBV29-1 TRBJ2-2 TRBD2
TRBV12-4,
CDK4R>L 4/5 58 CALPYGNKLVF TRAV38-1 TRAJ47 CASTPTGAYEQYF TRBJ2-7 TRBD2
TRBV12-3
TRBD1,
MLL2L>H 3/1 41 CALNPFNAGNMLTF TRAV12-2 TRAJ39 CSARVGDTGELFF TRBV20-1 TRBJ2-2
TRBD2
TRBD1,
MLL2L>H 3 / 19 45 CALNPFNAGNMLTF TRAV12-2 TRAJ39 CSARVGDTGELFF TRBV20-1 TRBJ2-2
TRBD2
TRBD1,
MLL2L>H 3 / 27 46 CALNPFNAGNMLTF TRAV12-2 TRAJ39 CSARVGDTGELFF TRBV20-1 TRBJ2-2
TRBD2
TRBD1,
MLL2L>H 3/5 42 CALNPFNAGNMLTF TRAV12-2 TRAJ39 CSARVGDTGELFF TRBV20-1 TRBJ2-2
TRBD2
TRBD1,
MLL2L>H 3/7 43 CALNPFNAGNMLTF TRAV12-2 TRAJ39 CSARVGDTGELFF TRBV20-1 TRBJ2-2
TRBD2
TRBD1,
MLL2L>H 3/9 44 CALNPFNAGNMLTF TRAV12-2 TRAJ39 CSARVGDTGELFF TRBV20-1 TRBJ2-2
TRBD2
MLL2L>H 4/1 47 CAVGEAGGGNKLTF TRAV8-3 TRAJ10 CASSYDSTTGELFF TRBV4-1 TRBJ2-2 TRBD1

MLL2L>H 4 / 10 50 CAVGEAGGGNKLTF TRAV8-3 TRAJ10 CASSYDSTTGELFF TRBV4-1 TRBJ2-2 TRBD1

MLL2L>H 4/5 48 CAVGEAGGGNKLTF TRAV8-3 TRAJ10 CASSYDSTTGELFF TRBV4-1 TRBJ2-2 TRBD1

MLL2L>H 4/6 49 CAVGEAGGGNKLTF TRAV8-3 TRAJ10 CASSYDSTTGELFF TRBV4-1 TRBJ2-2 TRBD1

SMARCD3H>Y 2/1 67 CAFMQFDMRF TRAV38-1 TRAJ43 CASTPGGYEQYF TRBV2 TRBJ2-7 TRBD1

SMARCD3H>Y 2 / 11 68 CAFMQFDMRF TRAV38-1 TRAJ43 CASTPGGYEQYF TRBV2 TRBJ2-7 TRBD1

TRBD1,
SMARCD3H>Y 4 / 10 60 CAVDTNTDKLIF TRAV2 TRAJ34 CASSQGYEQYF TRBV11-2 TRBJ2-7
TRBD2
TRBD1,
SMARCD3H>Y 4 / 16 61 CAVDTNTDKLIF TRAV2 TRAJ34 CASSQGYEQYF TRBV11-2 TRBJ2-7
TRBD2
TRBD1,
SMARCD3H>Y 4/6 59 CAVDTNTDKLIF TRAV2 TRAJ34 CASSQGYEQYF TRBV11-2 TRBJ2-7
TRBD2

21
Table S5. Well-characterized T-cell epitopes used for validation of peptide-MHC stability assay

Predicted
t1/2 β2-
Amino acid binding affinity t1/2 HLA-A*02:01
Name/protein microglobulin Reference
sequence (nM) to HLA- (hours)
(hours)
A*02:01*
Flu NP265-273 (A*03:01) ILRGSVAHK 24732 ** ** M. DiBrino et al., J. Immunol.
151, 5930-5935 (1993)
MART-126-35 WT EAAGIGILTV 5322.6 3.4 4.2 Y. Kawakami et al., J. Exp. Med.
180, 347-352 (1994)
MART-126-35 A27L ELAGIGILTV 253.9 14.5 16.8 D. Valmori et al., J. Immunol.
160, 1750-1758 (1998)
Flu MP58-66 (A*02:01) GILGFVFTL 15.7 25.5 28.9 J. Morrison et al., Eur. J.
Immunol. 22, 903-907 (1992)
gp100280-288 A288V YLEPGPVTV 17 23.6 27.5 M. R. Parkhurst et al., J.
Immunol. 157, 2539-2548 (1996)
CD20186-194 SLFLGILSV 7.4 20.2 20.9 J. Bae, J. A. Martinson, H. G.
Klingemann, Clin. Cancer Res.
11, 1629-1638 (2005)
CMV pp65495-503 NLVPMVATV 25.9 13.3 16.6 D. J. Diamond, J. York, J. Y. Sun,
C. L. Wright, S. J. Forman, Blood
90, 1751-1767 (1997)
*Predicted with NetMHC 4.0.
**Peptide does not bind HLA-A*02:01.

22
Table S6. HLA types of patients and healthy donors

Patient/ HLA-A HLA-B HLA-C HLA-DRB1

Donor
Allele 1 Allele 2 Allele 1 Allele 2 Allele 1 Allele 2 Allele 1 Allele 2
Patient 1 *02:01 *02:01 *15:01 *44:02 *03:04 *05:01
Donor 1 *01:AHVDJ *02:01 *08:YETY *44:AAAVY *03:AHVSP 07:AHVTD *03:01:01 *16:01:01
Donor 2 *02:01 *03:ABHBF *07:TXXS *40:AFXSK *03:04:01 *03:04:01 *08:01 *15:01:01
Donor 3 *01:AHVDJ *02:01 *39:06:02 *57:AGXDH *06:ADCTS *07:ADTKM *07:01:01 *15:01:01
Donor 4 *02:01 *02:01 *07:TXXS *55:01 *03:AHNNK *07:AEADZ *13:01:01 *15:01:01

Patient 2 *02:01 *11:01 *07:02 *35:01 *04:01 *07:02

Donor 5 *02:01 *02:01 *15:ACMGP *15:ACMGP *01:AGCRZ *03:AGDVM *04:01:01 *08:01
Donor 6 *02:01 *32:TKNC *14:AJ *15:ACMGP *04:ABNRU *08:AAPSN *07:01:01 *10:01:01
Donor 7 *02:01 *03:ABHBF *35:AESRM *44:AAAVY *04:ABMNU *05:XREK *04:01:01 *11:04:01
Donor 8 *02:01 *02:01 *15:AHFTX *27:VSZE *02:AHYKA *03:AJBAC *01:01:01 *08:01

HLA-A NMDP codes

*01:AHVDJ 01:01/01:14/01:26/01:30/01:66/01:98/01:100/01:104
*03:ABHBF 03:01/03:05/03:07/03:08/03:09/03:17/03:23/03:95/03:108/03:123/03:157/03:176
*32:TKNC 32:01/32:06/32:08/32:28/32:31/32:39

HLA-B NMDP codes

*07:TXXS 07:02/07:44/07:49N/07:58/07:59/07:61/07:120/07:128/07:129/07:130/07:156/07:161N/07:169
*08:YETY 08:01/08:19N/08:109
*14:AJ 14:01/14:08
*15:ACMGP 15:01/15:102/15:104/15:140/15:146/15:201/15:227/15:228/15:247/15:320/15:321Q
*15:AHFTX 15:01/15:43/15:102/15:104/15:117/15:129/15:140/15:146/15:201/15:225/15:227/15:228/15:247/15:259/15:320/15:321Q
*27:VSZE 27:05/27:08/27:13/27:37/27:38/27:67/27:70
*35:AESRM 35:03/35:70/35:279
*40:AFXSK 40:01/40:55/40:141/40:150/40:151/40:179/40:221/40:236/40:241/40:247/40:264/40:272/40:278/40:299/40:301
*44:AAAVY 44:02/44:02S/44:19N/44:27/44:66/44:118/44:187
*57:AGXDH 57:01/57:29/57:37/57:55/57:79N

HLA-C NMDP codes

*01:AGCRZ 01:02/01:22/01:48/01:51/01:65/01:107
*02:AHYKA 02:02/02:10/02:16/02:19/02:27/02:61/02:65
*03:AGDVM 03:03/03:04/03:18/03:20N/03:49/03:55/03:151/03:227
*03:AHNNK 03:03/03:04/03:11/03:13/03:18/03:20N/03:32/03:43/03:55/03:67/03:69/03:103/03:151/03:227/03:294
*03:AHVSP 03:04/03:08/03:09/03:32/03:35/03:37/03:38/03:41/03:64/03:92/03:162/03:279/03:294
*03:AJBAC 03:04/03:07/03:09/03:10/03:279/03:287
*04:ABMNU 04:01/04:10/04:15/04:33/04:103/04:120/04:129
*04:ABNRU 04:01/04:10/04:33/04:120/04:129
*05:XREK 05:01/05:09/05:25/05:29/05:46/05:64/05:78
*06:ADCTS 06:02/06:04/06:11/06:31/06:47/06:83/06:122/06:136
*07:ADTKM 07:02/07:27/07:31/07:49/07:76/07:138/07:241
*07:AEADZ 07:02/07:10/07:27/07:29/07:39/07:50/07:56/07:79/07:127/07:133/07:186/07:314/07:390
*07:AHVTD 07:01/07:16/07:20/07:27/07:40/07:153/07:196/07:212/07:248/07:257/07:263/07:304/07:330
*08:AAPSN 08:02/08:05/08:15/08:28/08:94

23
Table S7. Summary of exome sequencing data and potential neo-epitopes (separate file)

Table S8. All potential neo-epitopes from patient 1, 2 and 3 (separate file)

24
References
1. S. A. Rosenberg, N. P. Restifo, Adoptive cell transfer as personalized immunotherapy for
human cancer. Science 348, 62–68 (2015). Medline doi:10.1126/science.aaa4967
2. P. Sharma, J. P. Allison, The future of immune checkpoint therapy. Science 348, 56–61
(2015). Medline doi:10.1126/science.aaa8172
3. N. A. Rizvi, M. D. Hellmann, A. Snyder, P. Kvistborg, V. Makarov, J. J. Havel, W. Lee, J.
Yuan, P. Wong, T. S. Ho, M. L. Miller, N. Rekhtman, A. L. Moreira, F. Ibrahim, C.
Bruggeman, B. Gasmi, R. Zappasodi, Y. Maeda, C. Sander, E. B. Garon, T. Merghoub, J.
D. Wolchok, T. N. Schumacher, T. A. Chan, Mutational landscape determines sensitivity
to PD-1 blockade in non-small cell lung cancer. Science 348, 124–128 (2015). Medline
doi:10.1126/science.aaa1348
4. E. M. Van Allen, D. Miao, B. Schilling, S. A. Shukla, C. Blank, L. Zimmer, A. Sucker, U.
Hillen, M. H. Geukes Foppen, S. M. Goldinger, J. Utikal, J. C. Hassel, B. Weide, K. C.
Kaehler, C. Loquai, P. Mohr, R. Gutzmer, R. Dummer, S. Gabriel, C. J. Wu, D.
Schadendorf, L. A. Garraway, Genomic correlates of response to CTLA-4 blockade in
metastatic melanoma. Science 350, 207–211 (2015). Medline
doi:10.1126/science.aad0095
5. P. F. Robbins, Y. C. Lu, M. El-Gamil, Y. F. Li, C. Gross, J. Gartner, J. C. Lin, J. K. Teer, P.
Cliften, E. Tycksen, Y. Samuels, S. A. Rosenberg, Mining exomic sequencing data to
identify mutated antigens recognized by adoptively transferred tumor-reactive T cells.
Nat. Med. 19, 747–752 (2013). Medline doi:10.1038/nm.3161
6. N. van Rooij, M. M. van Buuren, D. Philips, A. Velds, M. Toebes, B. Heemskerk, L. J. van
Dijk, S. Behjati, H. Hilkmann, D. El Atmioui, M. Nieuwland, M. R. Stratton, R. M.
Kerkhoven, C. Kesmir, J. B. Haanen, P. Kvistborg, T. N. Schumacher, Tumor exome
analysis reveals neoantigen-specific T-cell reactivity in an ipilimumab-responsive
melanoma. J. Clin. Oncol. 31, e439–e442 (2013). Medline
doi:10.1200/JCO.2012.47.7521
7. C. Linnemann, M. M. van Buuren, L. Bies, E. M. Verdegaal, R. Schotte, J. J. Calis, S. Behjati,
A. Velds, H. Hilkmann, D. E. Atmioui, M. Visser, M. R. Stratton, J. B. Haanen, H. Spits,
S. H. van der Burg, T. N. Schumacher, High-throughput epitope discovery reveals
frequent recognition of neo-antigens by CD4+ T cells in human melanoma. Nat. Med. 21,
81–85 (2015). Medline doi:10.1038/nm.3773
8. E. Tran, M. Ahmadzadeh, Y. C. Lu, A. Gros, S. Turcotte, P. F. Robbins, J. J. Gartner, Z.
Zheng, Y. F. Li, S. Ray, J. R. Wunderlich, R. P. Somerville, S. A. Rosenberg,
Immunogenicity of somatic mutations in human gastrointestinal cancers. Science 350,
1387–1390 (2015). Medline doi:10.1126/science.aad1253
9. E. Tran, S. Turcotte, A. Gros, P. F. Robbins, Y. C. Lu, M. E. Dudley, J. R. Wunderlich, R. P.
Somerville, K. Hogan, C. S. Hinrichs, M. R. Parkhurst, J. C. Yang, S. A. Rosenberg,
Cancer immunotherapy based on mutation-specific CD4+ T cells in a patient with
epithelial cancer. Science 344, 641–645 (2014). Medline doi:10.1126/science.1251102
10. M. M. Gubin, X. Zhang, H. Schuster, E. Caron, J. P. Ward, T. Noguchi, Y. Ivanova, J.
Hundal, C. D. Arthur, W. J. Krebber, G. E. Mulder, M. Toebes, M. D. Vesely, S. S. Lam,
A. J. Korman, J. P. Allison, G. J. Freeman, A. H. Sharpe, E. L. Pearce, T. N.
Schumacher, R. Aebersold, H. G. Rammensee, C. J. Melief, E. R. Mardis, W. E.
Gillanders, M. N. Artyomov, R. D. Schreiber, Checkpoint blockade cancer
immunotherapy targets tumour-specific mutant antigens. Nature 515, 577–581 (2014).
Medline doi:10.1038/nature13988
11. H. Matsushita, M. D. Vesely, D. C. Koboldt, C. G. Rickert, R. Uppaluri, V. J. Magrini, C. D.
Arthur, J. M. White, Y. S. Chen, L. K. Shea, J. Hundal, M. C. Wendl, R. Demeter, T.
Wylie, J. P. Allison, M. J. Smyth, L. J. Old, E. R. Mardis, R. D. Schreiber, Cancer exome
analysis reveals a T-cell-dependent mechanism of cancer immunoediting. Nature 482,
400–404 (2012). Medline doi:10.1038/nature10755
12. M. Yadav, S. Jhunjhunwala, Q. T. Phung, P. Lupardus, J. Tanguay, S. Bumbaca, C. Franci,
T. K. Cheung, J. Fritsche, T. Weinschenk, Z. Modrusan, I. Mellman, J. R. Lill, L.
Delamarre, Predicting immunogenic tumour mutations by combining mass spectrometry
and exome sequencing. Nature 515, 572–576 (2014). Medline doi:10.1038/nature14001
13. D. Mittal, M. M. Gubin, R. D. Schreiber, M. J. Smyth, New insights into cancer
immunoediting and its three component phases—elimination, equilibrium and escape.
Curr. Opin. Immunol. 27, 16–25 (2014). Medline doi:10.1016/j.coi.2014.01.004
14. B. M. Carreno, V. Magrini, M. Becker-Hapak, S. Kaabinejadian, J. Hundal, A. A. Petti, A.
Ly, W. R. Lie, W. H. Hildebrand, E. R. Mardis, G. P. Linette, A dendritic cell vaccine
increases the breadth and diversity of melanoma neoantigen-specific T cells. Science 348,
803–808 (2015). Medline doi:10.1126/science.aaa3828
15. See supplementary materials on Science Online.
16. I. W. Abrahamsen, E. Stronen, S. Wälchli, J. N. Johansen, S. Kjellevoll, S. Kumari, M.
Komada, G. Gaudernack, G. Tjonnfjord, M. Toebes, T. N. Schumacher, F. Lund-
Johansen, J. Olweus, Targeting B cell leukemia with highly specific allogeneic T cells
with a public recognition motif. Leukemia 24, 1901–1909 (2010). Medline
doi:10.1038/leu.2010.186
17. T. Wölfel, M. Hauer, J. Schneider, M. Serrano, C. Wölfel, E. Klehmann-Hieb, E. De Plaen,
T. Hankeln, K. H. Meyer zum Büschenfelde, D. Beach, A p16INK4a-insensitive CDK4
mutant targeted by cytolytic T lymphocytes in a human melanoma. Science 269, 1281–
1284 (1995). Medline doi:10.1126/science.7652577
18. S. H. van der Burg, M. J. Visseren, R. M. Brandt, W. M. Kast, C. J. Melief, Immunogenicity
of peptides bound to MHC class I molecules depends on the MHC-peptide complex
stability. J. Immunol. 156, 3308–3314 (1996). Medline
19. M. Harndahl, M. Rasmussen, G. Roder, I. Dalgaard Pedersen, M. Sørensen, M. Nielsen, S.
Buus, Peptide-MHC class I stability is a better predictor than peptide affinity of CTL
immunogenicity. Eur. J. Immunol. 42, 1405–1416 (2012). Medline
doi:10.1002/eji.201141774
20. S. Kreiter, M. Vormehr, N. van de Roemer, M. Diken, M. Löwer, J. Diekmann, S. Boegel, B.
Schrörs, F. Vascotto, J. C. Castle, A. D. Tadmor, S. P. Schoenberger, C. Huber, Ö.
Türeci, U. Sahin, Mutant MHC class II epitopes drive therapeutic immune responses to
cancer. Nature 520, 692–696 (2015). Medline doi:10.1038/nature14426
21. C. A. Klebanoff, S. A. Rosenberg, N. P. Restifo, Prospects for gene-engineered T cell
immunotherapy for solid cancers. Nat. Med. 22, 26–36 (2016). Medline
doi:10.1038/nm.4015
22. G. Tosato, J. I. Cohen, Generation of Epstein-Barr Virus (EBV)-immortalized B cell lines.
Curr. Protoc. Immunol. Chapter 7, Unit 7 22 (2007).
23. H. Li, R. Durbin, Fast and accurate long-read alignment with Burrows-Wheeler transform.
Bioinformatics 26, 589–595 (2010). Medline doi:10.1093/bioinformatics/btp698
24. M. A. DePristo, E. Banks, R. Poplin, K. V. Garimella, J. R. Maguire, C. Hartl, A. A.
Philippakis, G. del Angel, M. A. Rivas, M. Hanna, A. McKenna, T. J. Fennell, A. M.
Kernytsky, A. Y. Sivachenko, K. Cibulskis, S. B. Gabriel, D. Altshuler, M. J. Daly, A
framework for variation discovery and genotyping using next-generation DNA
sequencing data. Nat. Genet. 43, 491–498 (2011). Medline doi:10.1038/ng.806
25. D. E. Larson, C. C. Harris, K. Chen, D. C. Koboldt, T. E. Abbott, D. J. Dooling, T. J. Ley, E.
R. Mardis, R. K. Wilson, L. Ding, SomaticSniper: Identification of somatic point
mutations in whole genome sequencing data. Bioinformatics 28, 311–317 (2012).
Medline doi:10.1093/bioinformatics/btr665
26. H. Li, B. Handsaker, A. Wysoker, T. Fennell, J. Ruan, N. Homer, G. Marth, G. Abecasis, R.
Durbin, The Sequence Alignment/Map format and SAMtools. Bioinformatics 25, 2078–
2079 (2009). Medline doi:10.1093/bioinformatics/btp352
27. P. Cingolani, A. Platts, L. Wang, M. Coon, T. Nguyen, L. Wang, S. J. Land, X. Lu, D. M.
Ruden, A program for annotating and predicting the effects of single nucleotide
polymorphisms, SnpEff: SNPs in the genome of Drosophila melanogaster strain w1118;
iso-2; iso-3. Fly 6, 80–92 (2012). Medline doi:10.4161/fly.19695
28. C. Trapnell, L. Pachter, S. L. Salzberg, TopHat: Discovering splice junctions with RNA-Seq.
Bioinformatics 25, 1105–1111 (2009). Medline doi:10.1093/bioinformatics/btp120
29. A. Roberts, C. Trapnell, J. Donaghey, J. L. Rinn, L. Pachter, Improving RNA-Seq expression
estimates by correcting for fragment bias. Genome Biol. 12, R22 (2011). Medline
doi:10.1186/gb-2011-12-3-r22
30. S. Behjati, P. S. Tarpey, H. Sheldon, I. Martincorena, P. Van Loo, G. Gundem, D. C. Wedge,
M. Ramakrishna, S. L. Cooke, N. Pillay, H. K. Vollan, E. Papaemmanuil, H. Koss, T. D.
Bunney, C. Hardy, O. R. Joseph, S. Martin, L. Mudie, A. Butler, J. W. Teague, M. Patil,
G. Steers, Y. Cao, C. Gumbs, D. Ingram, A. J. Lazar, L. Little, H. Mahadeshwar, A.
Protopopov, G. A. Al Sannaa, S. Seth, X. Song, J. Tang, J. Zhang, V. Ravi, K. E. Torres,
B. Khatri, D. Halai, I. Roxanis, D. Baumhoer, R. Tirabosco, M. F. Amary, C. Boshoff, U.
McDermott, M. Katan, M. R. Stratton, P. A. Futreal, A. M. Flanagan, A. Harris, P. J.
Campbell, Recurrent PTPRB and PLCG1 mutations in angiosarcoma. Nat. Genet. 46,
376–379 (2014). Medline doi:10.1038/ng.2921
31. M. Nielsen, C. Lundegaard, O. Lund, C. Keşmir, The role of the proteasome in generating
cytotoxic T-cell epitopes: Insights obtained from improved predictions of proteasomal
cleavage. Immunogenetics 57, 33–41 (2005). Medline doi:10.1007/s00251-005-0781-7
32. C. Lundegaard, K. Lamberth, M. Harndahl, S. Buus, O. Lund, M. Nielsen, NetMHC-3.0:
Accurate web accessible predictions of human, mouse and monkey MHC class I affinities
for peptides of length 8-11. Nucleic Acids Res. 36, W509–W512 (2008). Medline
doi:10.1093/nar/gkn202
33. M. Nielsen, C. Lundegaard, P. Worning, C. S. Hvid, K. Lamberth, S. Buus, S. Brunak, O.
Lund, Improved prediction of MHC class I and class II epitopes using a novel Gibbs
sampling approach. Bioinformatics 20, 1388–1397 (2004). Medline
doi:10.1093/bioinformatics/bth100
34. M. M. van Buuren, J. J. Calis, T. N. Schumacher, High sensitivity of cancer exome-based
CD8 T cell neo-antigen identification. OncoImmunology 3, e28836 (2014). Medline
doi:10.4161/onci.28836
35. S. Kumari, S. Wälchli, L. E. Fallang, W. Yang, F. Lund-Johansen, T. N. Schumacher, J.
Olweus, Alloreactive cytotoxic T cells provide means to decipher the immunopeptidome
and reveal a plethora of tumor-associated self-epitopes. Proc. Natl. Acad. Sci. U.S.A. 111,
403–408 (2014). Medline doi:10.1073/pnas.1306549111
36. M. Toebes, M. Coccoris, A. Bins, B. Rodenko, R. Gomez, N. J. Nieuwkoop, W. van de
Kasteele, G. F. Rimmelzwaan, J. B. Haanen, H. Ovaa, T. N. Schumacher, Design and use
of conditional MHC class I ligands. Nat. Med. 12, 246–251 (2006). Medline
doi:10.1038/nm1360
37. S. R. Hadrup, A. H. Bakker, C. J. Shu, R. S. Andersen, J. van Veluw, P. Hombrink, E.
Castermans, P. Thor Straten, C. Blank, J. B. Haanen, M. H. Heemskerk, T. N.
Schumacher, Parallel detection of antigen-specific T-cell responses by multidimensional
 encoding of MHC multimers. Nat. Methods 6, 520–526 (2009). Medline
doi:10.1038/nmeth.1345
38. F. Tang, C. Barbacioru, E. Nordman, B. Li, N. Xu, V. I. Bashkirov, K. Lao, M. A. Surani,
RNA-Seq analysis to capture the transcriptome landscape of a single cell. Nat. Protoc. 5,
516–535 (2010). Medline doi:10.1038/nprot.2009.236
39. C. Linnemann, B. Heemskerk, P. Kvistborg, R. J. Kluin, D. A. Bolotin, X. Chen, K. Bresser,
M. Nieuwland, R. Schotte, S. Michels, R. Gomez-Eerland, L. Jahn, P. Hombrink, N.
Legrand, C. J. Shu, I. Z. Mamedov, A. Velds, C. U. Blank, J. B. Haanen, M. A.
Turchaninova, R. M. Kerkhoven, H. Spits, S. R. Hadrup, M. H. Heemskerk, T.
Blankenstein, D. M. Chudakov, G. M. Bendle, T. N. Schumacher, High-throughput
identification of antigen-specific TCRs by TCR gene capture. Nat. Med. 19, 1534–1541
(2013). Medline doi:10.1038/nm.3359
40. X. Brochet, M. P. Lefranc, V. Giudicelli, IMGT/V-QUEST: The highly customized and
integrated system for IG and TR standardized V-J and V-D-J sequence analysis. Nucleic
Acids Res. 36, W503–W508 (2008). Medline doi:10.1093/nar/gkn316
41. F. A. Ran, P. D. Hsu, J. Wright, V. Agarwala, D. A. Scott, F. Zhang, Genome engineering
using the CRISPR-Cas9 system. Nat. Protoc. 8, 2281–2308 (2013). Medline
doi:10.1038/nprot.2013.143
42. W. Wu, H. Slåstad, D. de la Rosa Carrillo, T. Frey, G. Tjønnfjord, E. Boretti, H. C. Aasheim,
V. Horejsi, F. Lund-Johansen, Antibody array analysis with label-based detection and
resolution of protein size. Mol. Cell. Proteomics 8, 245–257 (2009). Medline
doi:10.1074/mcp.M800171-MCP200
43. X. Robin, N. Turck, A. Hainard, N. Tiberti, F. Lisacek, J. C. Sanchez, M. Müller, pROC: An
open-source package for R and S+ to analyze and compare ROC curves. BMC
Bioinformatics 12, 77 (2011). Medline doi:10.1186/1471-2105-12-77