Datasets were obtained from the GEO database through the following GEO Series accession numbers: GSE93391, GSE114012, GSE131594, GSE152699, GSE124854, GSE135215, GSE99116, GSE124109, GSE61130, GSE64553 and GSE63577.
Cell lines were obtained from ATCC or Sigma and regularly checked for mycoplasma. A549 and NCIH460 were cultured in DMEM . NCIH358, NCIH1299 and NCIH1563 were maintained in RPMI-1640 supplemented with 5 mM sodium pyruvate and 0.5% glucose. NCIH1944, NCIH1666, NCIH1650 and L23 were grown in RPMI-1640 ATCC formulation . A427 were cultured in EMEM . A549, NCIH460, H358, NCIH1299, NCIH1563, and A427 were supplemented with 10% heat inactivated FBS.
Endogenous PCNA was labelled at the N-terminus with a cDNA encoding mRuby in both A549 and NCI-H1944 cells, using AAV-mediated gene-targeting, according to methods described in Zerjatke et al. []. mRuby-expressing cells were sorted into 50:50 conditioned:fresh media at single-cell density into 96-well plates by FACS and single-cell clones expanded. For live-cell imaging, 500 cells in phenol-red free media were plated per well of a 384 CellCarrierUltra plate the day before imaging.
The G0 arrest scores for cancer cell lines were calculated using corresponding log-transformed RPKM normalised bulk RNA-seq data from the Cancer Cell Line Encyclopedia database [CEP89 was depleted by ON-Target siRNA Pool from Horizon. NCI-H1299 cells were reverse transfected in 384-well plates with 20 nM of non-targeting control or CEP89-targeting siRNA using Lipofectamine RNAiMax , according to the manufacturer’s instructions.
The mutation rates of all TCGA primary tumour samples were determined by log-transforming the total number of mutations in each sample divided by the length of the exome capture .functional status was assessed based on somatic mutation and copy number alterations as described in Zhang et al. [