13The set of proteins required for mitotic division remains poorly characterised. Here, an 14 extensive series of correlation analyses of human and mouse transcriptomics data was 15 performed to identify genes strongly and reproducibly associated with cells undergoing S/G2-16M phases of the cell cycle. In so doing, a list of 701 cell cycle-associated genes was defined 17and shown that whilst many are only expressed during these phases, the expression of others 18 is also driven by alternative promoters. Of this list, 496 genes have known cell cycle functions, 19whereas 205 were assigned as putative cell cycle genes, 53 of which are functionally 20uncharacterised. Among these, 27 were screened for subcellular localisation revealing many 21to be nuclear localised and at least four to be novel centrosomal proteins. Furthermore, 10 22 others inhibited cell proliferation upon siRNA knockdown. This study presents the first 23comprehensive list of human cell cycle proteins, identifying many new candidate proteins. 24 33 are important therapeutic targets 3 and novel components might present new therapeutic 34opportunities.
36Many of the key proteins required for mitotic division are known from studies in model 37 organisms including yeast, as well as mammalian cells 4,5 . With the aim of identifying all the 38 components of the system, high content analysis platforms have been utilised. For example 39RNAi screens 6-8 , CRISPR/Cas9 9 , proteomics 10,11,12 studies have all proposed lists of cell cycle 40 genes/proteins but a consensus between studies has not emerged. In particular, genome-wide 41transcriptomics studies [13][14][15][16][17][18][19] have identified sets of transcripts sequentially regulated during 42 the cell cycle phases in multiple species but comparison of results from four studies of 43 different human cell lines identified only 96 genes in common 15 . Our reanalysis of these data, 44taking account of some of the technical variables, suggested that the true concordance of the 45 cell cycle transcriptional network across cell types is much greater 20 . Furthermore, our 46analyses of large collections of tissue and cell transcriptomics data commonly identify a large 47 cluster of cell cycle transcripts whose expression is elevated in cells or tissues with a high 48 mitotic index [21][22][23][24][25] . 49 50We report here on a data driven curation exercise aimed at identifying the cohort of genes 51up-regulated in all human cell types from the G1/S boundary through to the completion of M 52 phase (S/G2-M). There are of course, many growth-associated transcriptional regulatory 53 events including activation of cyclins, cyclin-dependent kinases and E2F transcription factors, 54 that occur during G1 and are a precondition for entry into S phase 26 , but these are not the 55 focus of this study. We monitored genome-wide gene expression in primary human dermal 56fibroblasts (NHDF) cells as they synchronously enter the cell cycle from a resting state (G0). 57Using network co-expression analysis and clustering of ...