Erratum: Core transcription programs controlling injury-induced neurodegeneration of retinal ganglion cells (Neuron (2024) 112(14) (2452), (S0896627324004458), (10.1016/j.neuron.2024.06.010))

(Neuron 110, 2607–2624.e1–e8; August 17, 2022) In our published article, we mistakenly used an incorrect intermediate GEO file GSE190667_rna.DEX to generate Tables S5 and S6 and Figures 4, 5, and S4. This file did not have the complete set of eventual covariates used, nor complete outlier removal, and was based on a false discovery rate (FDR) threshold of <0.2, which we used for preliminary analysis rather than the more stringent FDR < 0.1 that we preferred for definitive analyses. The final outliers include ATF4.3d.6, ATF4.3d.7, ATF4.3d.5, CEBPG.3d.7, non.3d.9, and non.3d.11. We previously omitted “AlignSeq.PC2” from the regression model, and the correct regression model is ∼ Genotype + AlignSeq.PC1 + AlignSeq.PC2 + AlignSeq.PC3 + AlignSeq.PC4. We have made the code to reproduce these analyses available on GitHub (10.5281/zenodo.10878059). Based on the corrected GSE190667_rna.DEX (https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE190667), corrected Figures 4, S4, and 5 are provided below, and the supplemental tables have been corrected online. Due to the more stringent criteria used in the corrected analysis to identify differentially expressed genes (DEGs), approximately 20%–30% of the original gene targets were lost. As a result, the Gene Ontology (GO) terms “death receptor signaling” and “cardiac hypertrophy signaling” are no longer significant and are removed in the corrected Figure 4. Although the other GO terms remain, their order has shifted, reflecting slight differences in p values. The representative genes Tnfrsf21 and death receptor 6 (DR6) are also lost. Accordingly, the following concluding statements from the results section, along with the mentioned references, no longer apply: “For example, Tnfrsf21 or death receptor 6 (DR6), a TNF receptor family member that triggers cell death (Benschop et al., 2009), is uniquely bound and activated by ATF3 and CHOP, but not C/EBPγ and ATF4 (Figures 5A and 5B, ATF3/CHOP positive; Figure S5). In contrast, Hdac9, which regulates endogenous DNA repair (Wong et al., 2009), is a unique target of C/EBPγ and ATF4 (Figures 5A and 5B, C/EBPγ/ATF4 positive; Figure S5).” While changing a subset of DEGs, these corrections do not affect the main findings or conclusions of the manuscript. The authors would like to apologize for any inconvenience caused.