A causality-based machine learning Pipeline for Cancer Inference (PiCnIc) is introduced to infer the underlying somatic evolution of ensembles of tumors from next-generation sequencing data. PiCnIc combines techniques for sample stratification, driver selection, and identification of fitness-equivalent exclusive alterations to exploit an algorithm based on Suppes’ probabilistic causation. The accuracy and translational significance of the results are studied in detail, with an application to colorectal cancer. The PiCnIc pipeline has been made publicly accessible for reproducibility, interoperability, and future enhancements.

Algorithmic methods to infer the evolutionary trajectories in cancer progression

Caravagna G
;
2016-01-01

Abstract

A causality-based machine learning Pipeline for Cancer Inference (PiCnIc) is introduced to infer the underlying somatic evolution of ensembles of tumors from next-generation sequencing data. PiCnIc combines techniques for sample stratification, driver selection, and identification of fitness-equivalent exclusive alterations to exploit an algorithm based on Suppes’ probabilistic causation. The accuracy and translational significance of the results are studied in detail, with an application to colorectal cancer. The PiCnIc pipeline has been made publicly accessible for reproducibility, interoperability, and future enhancements.
Pubblicato
https://www.pnas.org/content/113/28/E4025
File in questo prodotto:
File Dimensione Formato  
PNAS-2016-Caravagna-E4025-34.pdf

accesso aperto

Tipologia: Documento in Versione Editoriale
Licenza: Creative commons
Dimensione 2.45 MB
Formato Adobe PDF
2.45 MB Adobe PDF Visualizza/Apri

I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.

Utilizza questo identificativo per citare o creare un link a questo documento: https://hdl.handle.net/11368/2956278
Citazioni
  • ???jsp.display-item.citation.pmc??? ND
  • Scopus 51
  • ???jsp.display-item.citation.isi??? 44
social impact