Background: Despite low rates of local recurrence in primary uveal melanomas following radical treatment, approximately half of all patients will develop metastatic disease. There are limited effective systemic therapeutic options and survival outcomes remain poor. Genome-wide CRISPR-Cas9 knockout screening enables appropriate molecular targets and pathways that underlie cancer development to be identified. Methods: Genome-wide single gRNA CRISPR knockout screens were performed using Human Improved Genome-wide Knockout CRISPR Library v1.1. Eight human uveal melanoma cell lines were screened for 14 days to identify gene dependencies. Prior to screening, all cell lines underwent RNA-Seq to quantify gene expression. CRISPR screen hits were validated using single gRNA competitive growth assays. In vitro functional assessment of CDS2 knockout was performed using inducible CDS2 knockout isogenic cell lines in colony forming assays and Annexin V apoptosis assays. To determine the effects of CDS2 loss on phosphoinositide synthesis, quantification of phosphatidic acid and phosphoinositides was performed using HPLC-MS. To test the association between CDS2 essentiality and CDS1 expression across all cancer types, we analysed gene essentiality data from Project Score and Project Achilles with gene expression data obtained from Cell Model Passports and the DepMap download portal (22Q4). Cell lines with CDS1 expression log2(TPM+1) < 1 were annotated as low-expressors, and the rest were annotated as high-expressors. Single group data were assessed using Student's t test. Results: We identified CDS2, which encodes enzymes that maintain phosphoinositide levels during PLC signalling, to be an essential gene in 7 out of 8 screened cell lines. All cell lines were found to have low expression of its paralogous gene, CDS1. Doxycycline-inducible CDS2 knockout clones were generated from cell lines MP41 and OMM2.5. Reduced colony formation, with a significant decrease in surviving fraction was observed after doxycycline compared with treatment with a DMSO control in both cell lines (P < 0.0001). Apoptotic and dead cells increased significantly when doxycycline was added in MP41(13.6 ± 0.9% to 47.4 ± 3.3%; P= 0.0006) and OMM2.5 (3.7 ± 0.4% to 60.4± 0.7%, P< 0.0001). CDS2 knockout resulted in an increase in the precursor phosphatidic acid, and a significant reduction of the phosphoinositides PI (MP41, P= 0.018; OMM2.5, P= 0.0005) and PIP (MP41, P= 0.033; OMM2.5, P= 0.006). CDS2 was more essential in CDS1 low-expressors in both Project Score (P< 0.0001) and Project Achilles (P< 0.0001) datasets, compared with CDS1 high-expressors. Conclusions: Disruption of CDS2 results in reduced cellular fitness in uveal melanoma. Given the synthetic lethal relationship between CDS1 and CDS2, in the context of low CDS1 gene expression as observed in uveal melanomas and in other tumour types, CDS2 is a potential novel therapeutic target.