“…In this study, the authors determined a gene signature for FOLFOX based on the gene expression of 14 classifier genes. In more detail, the 14-gene signature was constituted by different genes, including the smad ubiquitin regulatory factor 2 ( SMURF2 ), a negative regulator of TGF-β signaling with a tumor-suppressive role [ 107 ], the Mbt domain containing 1 ( MBTD1 ), encoding a potent transcriptional coactivator [ 108 ], the adaptor related protein complex 3 subunit Mu 2 ( AP3M2 ), the RING finger protein 141 ( RNF141 ), a gene reported to interact with KRAS promoting CRC progression [ 109 ], the aminopeptidase puromycin sensitive ( NPEPPS ), encoding key zinc metallopeptidases [ 110 ], the bromodomain PHD finger transcription factor ( BPTF ), a chromatin remodeling-related gene [ 111 ], the family with sequence similarity 73, member A ( FAM73A ), the amyloid beta precursor protein binding protein 2 ( APPBP2 ), the archaelysin family metallopeptidase 2 pseudogene 1 ( AMZ2P1 ), the SLIT-ROBO rho GTPase activating protein 1 ( SRGAP1 ), that encodes a protein mediating cell migration associated with CRC tumor progression and poor prognosis [ 112 ], the N-myristoyltransferase-1 ( NMT1 ), necessary for lysosomal degradation and mammalian target of rapamycin (mTOR) signaling pathway [ 113 ], the centrosome and spindle pole associated protein 1 ( CSPP1 ), whose circ-CSPP1 was found to be significantly overexpressed in CRC tissues [ 114 ], the eukaryotic translation initiation factor 1 ( EIF1 ), and the centrosomal protein 290 ( CEP290 ) genes. Using the 14- gene classifier, the learning model was capable of classifying responders with an overall SE of 91.3% compared with non-responders with 95.6% in SP, achieving an ACC of 80.2%.…”