PMID- 37420249
OWN - NLM
STAT- MEDLINE
DCOM- 20230721
LR  - 20231116
IS  - 1756-994X (Electronic)
IS  - 1756-994X (Linking)
VI  - 15
IP  - 1
DP  - 2023 Jul 7
TI  - Mutation-Attention (MuAt): deep representation learning of somatic mutations for 
      tumour typing and subtyping.
PG  - 47
LID - 10.1186/s13073-023-01204-4 [doi]
LID - 47
AB  - BACKGROUND: Cancer genome sequencing enables accurate classification of tumours 
      and tumour subtypes. However, prediction performance is still limited using 
      exome-only sequencing and for tumour types with low somatic mutation burden such 
      as many paediatric tumours. Moreover, the ability to leverage deep representation 
      learning in discovery of tumour entities remains unknown. METHODS: We introduce 
      here Mutation-Attention (MuAt), a deep neural network to learn representations of 
      simple and complex somatic alterations for prediction of tumour types and 
      subtypes. In contrast to many previous methods, MuAt utilizes the attention 
      mechanism on individual mutations instead of aggregated mutation counts. RESULTS: 
      We trained MuAt models on 2587 whole cancer genomes (24 tumour types) from the 
      Pan-Cancer Analysis of Whole Genomes (PCAWG) and 7352 cancer exomes (20 types) 
      from the Cancer Genome Atlas (TCGA). MuAt achieved prediction accuracy of 89% for 
      whole genomes and 64% for whole exomes, and a top-5 accuracy of 97% and 90%, 
      respectively. MuAt models were found to be well-calibrated and perform well in 
      three independent whole cancer genome cohorts with 10,361 tumours in total. We 
      show MuAt to be able to learn clinically and biologically relevant tumour 
      entities including acral melanoma, SHH-activated medulloblastoma, SPOP-associated 
      prostate cancer, microsatellite instability, POLE proofreading deficiency, and 
      MUTYH-associated pancreatic endocrine tumours without these tumour subtypes and 
      subgroups being provided as training labels. Finally, scrunity of MuAt attention 
      matrices revealed both ubiquitous and tumour-type specific patterns of simple and 
      complex somatic mutations. CONCLUSIONS: Integrated representations of somatic 
      alterations learnt by MuAt were able to accurately identify histological tumour 
      types and identify tumour entities, with potential to impact precision cancer 
      medicine.
CI  - (c) 2023. The Author(s).
FAU - Sanjaya, Prima
AU  - Sanjaya P
AD  - Institute for Molecular Medicine Finland (FIMM), HiLIFE, University of Helsinki, 
      Helsinki, Finland.
AD  - Applied Tumor Genomics Research Program, Faculty of Medicine, University of 
      Helsinki, Helsinki, Finland.
AD  - iCAN Digital Precision Cancer Medicine Flagship, Helsinki, Finland.
FAU - Maljanen, Katri
AU  - Maljanen K
AD  - Institute for Molecular Medicine Finland (FIMM), HiLIFE, University of Helsinki, 
      Helsinki, Finland.
AD  - Applied Tumor Genomics Research Program, Faculty of Medicine, University of 
      Helsinki, Helsinki, Finland.
AD  - iCAN Digital Precision Cancer Medicine Flagship, Helsinki, Finland.
FAU - Katainen, Riku
AU  - Katainen R
AD  - Institute for Molecular Medicine Finland (FIMM), HiLIFE, University of Helsinki, 
      Helsinki, Finland.
AD  - Applied Tumor Genomics Research Program, Faculty of Medicine, University of 
      Helsinki, Helsinki, Finland.
AD  - iCAN Digital Precision Cancer Medicine Flagship, Helsinki, Finland.
AD  - Department of Medical and Clinical Genetics, Faculty of Medicine, University of 
      Helsinki, Helsinki, Finland.
FAU - Waszak, Sebastian M
AU  - Waszak SM
AD  - Centre for Molecular Medicine Norway (NCMM), Nordic EMBL Partnership, University 
      of Oslo and Oslo University Hospital, Oslo, Norway.
AD  - Swiss Institute for Experimental Cancer Research School of Life Sciences, Ecole 
      Polytechnique Federale de Lausanne (EPFL), Lausanne, Switzerland.
AD  - Department of Neurology, University of California, San Francisco (UCSF), San 
      Francisco, CA, USA.
CN  - Genomics England Research Consortium
FAU - Aaltonen, Lauri A
AU  - Aaltonen LA
AD  - Applied Tumor Genomics Research Program, Faculty of Medicine, University of 
      Helsinki, Helsinki, Finland.
AD  - Department of Medical and Clinical Genetics, Faculty of Medicine, University of 
      Helsinki, Helsinki, Finland.
FAU - Stegle, Oliver
AU  - Stegle O
AD  - Division of Computational Genomics and Systems Genetics, German Cancer Research 
      Center (DKFZ), Heidelberg, Germany.
AD  - Genome Biology Unit, European Molecular Biology Laboratory, Heidelberg, Germany.
FAU - Korbel, Jan O
AU  - Korbel JO
AD  - Division of Computational Genomics and Systems Genetics, German Cancer Research 
      Center (DKFZ), Heidelberg, Germany.
AD  - Genome Biology Unit, European Molecular Biology Laboratory, Heidelberg, Germany.
AD  - European Molecular Biology Laboratory, European Bioinformatics Institute, 
      Wellcome Genome Campus, Hinxton, Cambridge, UK.
FAU - Pitkanen, Esa
AU  - Pitkanen E
AUID- ORCID: 0000-0002-9818-6370
AD  - Institute for Molecular Medicine Finland (FIMM), HiLIFE, University of Helsinki, 
      Helsinki, Finland. esa.pitkanen@helsinki.fi.
AD  - Applied Tumor Genomics Research Program, Faculty of Medicine, University of 
      Helsinki, Helsinki, Finland. esa.pitkanen@helsinki.fi.
AD  - iCAN Digital Precision Cancer Medicine Flagship, Helsinki, Finland. 
      esa.pitkanen@helsinki.fi.
AD  - Genome Biology Unit, European Molecular Biology Laboratory, Heidelberg, Germany. 
      esa.pitkanen@helsinki.fi.
LA  - eng
GR  - 322675/Academy of Finland/
GR  - 328890/Academy of Finland/
GR  - 187615/Norges Forskningsrad/
PT  - Journal Article
DEP - 20230707
PL  - England
TA  - Genome Med
JT  - Genome medicine
JID - 101475844
SB  - IM
MH  - *Neoplasms/genetics/pathology
MH  - Humans
MH  - Deep Learning
MH  - Benchmarking
MH  - *Mutation
PMC - PMC10326961
OTO - NOTNLM
OT  - Cancer genomics
OT  - Deep learning
OT  - Deep neural networks
OT  - Machine learning
OT  - Molecular tumour subtyping
OT  - Precision cancer medicine
OT  - Somatic mutations
OT  - Tumour type prediction
OT  - Whole exome sequencing
OT  - Whole genome sequencing
COIS- The authors declare that they have no competing interests.
FIR - Ambrose, J C
IR  - Ambrose JC
FIR - Arumugam, P
IR  - Arumugam P
FIR - Bevers, R
IR  - Bevers R
FIR - Bleda, M
IR  - Bleda M
FIR - Boardman-Pretty, F
IR  - Boardman-Pretty F
FIR - Boustred, C R
IR  - Boustred CR
FIR - Brittain, H
IR  - Brittain H
FIR - Brown, M A
IR  - Brown MA
FIR - Caulfield, M J
IR  - Caulfield MJ
FIR - Chan, G C
IR  - Chan GC
FIR - Giess, A
IR  - Giess A
FIR - Griffin, J N
IR  - Griffin JN
FIR - Hamblin, A
IR  - Hamblin A
FIR - Henderson, S
IR  - Henderson S
FIR - Hubbard, T J P
IR  - Hubbard TJP
FIR - Jackson, R
IR  - Jackson R
FIR - Jones, L J
IR  - Jones LJ
FIR - Kasperaviciute, D
IR  - Kasperaviciute D
FIR - Kayikci, M
IR  - Kayikci M
FIR - Kousathanas, A
IR  - Kousathanas A
FIR - Lahnstein, L
IR  - Lahnstein L
FIR - Lakey, A
IR  - Lakey A
FIR - Leigh, S E A
IR  - Leigh SEA
FIR - Leong, I U S
IR  - Leong IUS
FIR - Leong, F J
IR  - Leong FJ
FIR - Maleady-Crowe, F
IR  - Maleady-Crowe F
FIR - McEntagart, M
IR  - McEntagart M
FIR - Minneci, F
IR  - Minneci F
FIR - Mitchell, J
IR  - Mitchell J
FIR - Moutsianas, L
IR  - Moutsianas L
FIR - Mueller, M
IR  - Mueller M
FIR - Murugaesu, N
IR  - Murugaesu N
FIR - Need, A C
IR  - Need AC
FIR - O'Donovan, P
IR  - O'Donovan P
FIR - Odhams, C A
IR  - Odhams CA
FIR - Patch, C
IR  - Patch C
FIR - Perez-Gil, D
IR  - Perez-Gil D
FIR - Perez-Gil, M B
IR  - Perez-Gil MB
FIR - Pullinger, J
IR  - Pullinger J
FIR - Rahim, T
IR  - Rahim T
FIR - Rendon, A
IR  - Rendon A
FIR - Rogers, T
IR  - Rogers T
FIR - Savage, K
IR  - Savage K
FIR - Sawant, K
IR  - Sawant K
FIR - Scott, R H
IR  - Scott RH
FIR - Siddiq, A
IR  - Siddiq A
FIR - Siddiq, A
IR  - Siddiq A
FIR - Smith, S C
IR  - Smith SC
FIR - Sosinsky, A
IR  - Sosinsky A
FIR - Stuckey, A
IR  - Stuckey A
FIR - Tanguy, M
IR  - Tanguy M
FIR - Taylor Tavares, A L
IR  - Taylor Tavares AL
FIR - Thomas, E R A
IR  - Thomas ERA
FIR - Thompson, S R
IR  - Thompson SR
FIR - Tucci, A
IR  - Tucci A
FIR - Welland, M J
IR  - Welland MJ
FIR - Williams, E
IR  - Williams E
FIR - Witkowska, K
IR  - Witkowska K
FIR - Wood, S M
IR  - Wood SM
FIR - Zarowiecki, M
IR  - Zarowiecki M
EDAT- 2023/07/08 10:42
MHDA- 2023/07/10 06:42
PMCR- 2023/07/07
CRDT- 2023/07/07 23:38
PHST- 2022/06/19 00:00 [received]
PHST- 2023/06/21 00:00 [accepted]
PHST- 2023/07/10 06:42 [medline]
PHST- 2023/07/08 10:42 [pubmed]
PHST- 2023/07/07 23:38 [entrez]
PHST- 2023/07/07 00:00 [pmc-release]
AID - 10.1186/s13073-023-01204-4 [pii]
AID - 1204 [pii]
AID - 10.1186/s13073-023-01204-4 [doi]
PST - epublish
SO  - Genome Med. 2023 Jul 7;15(1):47. doi: 10.1186/s13073-023-01204-4.