ABSTRACTDespite the ubiquity and ecological importance of diatoms, much remains to be understood about their physiology and metabolism, including their carotenoid biosynthesis pathway. Early carotenoid biosynthesis steps are well-conserved, while the identity of the enzymes that catalyze the later steps and their order remain unclear. Those steps lead to the biosynthesis of the final pathway products: the main accessory light-harvesting pigment fucoxanthin (Fx) and the main photoprotective pigment pool comprised of diadinoxanthin (Ddx) and its reversibly de-epoxidized form diatoxanthin (Dtx). We used sequence comparison to known carotenoid biosynthesis enzymes to identify novel candidates in the diatom Thalassiosira pseudonana. Microarray and RNA-seq data was used to select candidates with transcriptomic responses similar to known carotenoid biosynthesis genes and to create full-length gene models, and we focused on those that encode proteins predicted to be chloroplast-localized. We identified a violaxanthin de-epoxidase-like gene (Thaps3_11707, VDL2) that when overexpressed results in increased Fx abundance while stoichiometrically reducing Ddx+Dtx. Based on transcriptomics, we hypothesize that Thaps3_10233 may also contribute to Fx biosynthesis, in addition to VDL2. Separately using antisense RNA to target VDL2, VDL1, and both LUT1-like copies (hypothesized to catalyze an earlier step in the pathway) simultaneously, reduced the overall cellular photosynthetic pigment content, including chlorophylls, suggesting destabilization of light-harvesting complexes by Fx deficiency. Based on transcriptomic and physiological data, we hypothesize that the two predicted T. pseudonana zeaxanthin epoxidases have distinct functions and that different copies of phytoene synthase and phytoene desaturase may serve to initiate carotenoid biosynthesis in response to different cellular needs. Finally, nine carotene cis/trans isomerase (CRTISO) candidates identified based on sequence identity to known CRTISO proteins were narrowed to two most likely to be part of the T. pseudonana carotenoid biosynthesis pathway based on transcriptomic responses and predicted chloroplast targeting.