Mutations affecting the mitochondrial fusion protein Optic Atrophy 1 (OPA1) cause autosomal dominant optic atrophy (DOA) – one of the most common form of mitochondrial disease. The majority of patients develop isolated optic atrophy, but about 20% of OPA1 mutation carriers manifest more severe neurological deficits as part of a “DOA+” phenotype. OPA1 deficiency causes mitochondrial fragmentation and also disrupts cristae organization, oxidative phosphorylation, mitochondrial DNA (mtDNA) maintenance, and cell viability. It has not yet been established whether phenotypic severity can be modulated by genetic modifiers of OPA1. To better understand the genetic regulation of mitochondrial dynamics, we established a high-throughput imaging pipeline using supervised machine learning (ML) to perform unbiased, quantitative mitochondrial morphology analysis that was coupled with a bespoke siRNA library targeting the entire known mitochondrial proteome (1531 genes), providing a detailed phenotypic screening of human fibroblasts. In control fibroblasts, we identified known and novel genes whose depletion promoted elongation or fragmentation of the mitochondrial network. In DOA+ patient fibroblasts, we identified 91 candidate genes whose depletion prevents mitochondrial fragmentation, including the mitochondrial fission genes DNM1L, MIEF1, and SLC25A46, but also genes not previously linked to mitochondrial dynamics such as Phosphatidyl Glycerophosphate Synthase (PGS1), which belongs to the cardiolipin (CL) synthesis pathway. PGS1 depletion reduces CL content in mitochondria and rebalances mitochondrial dynamics in OPA1-deficient fibroblasts by inhibiting mitochondrial fission, which improves defective respiration, but does not rescue mtDNA depletion, cristae dysmorphology or apoptotic sensitivity. Our data reveal that the multifaceted roles of OPA1 in mitochondria can be functionally uncoupled by modulating mitochondrial lipid metabolism, providing novel insights into the cellular relevance of mitochondrial fragmentation. This study illustrates the power of a first-in-kind objective automated imaging approach to uncover genetic modifiers of mitochondrial disease through high-throughput phenotypic screening of patient fibroblasts.