We note several limitations. First, we again emphasize that we can implicate a particular cell type but it is premature to exclude cell types for which we do not have data 12. Second, we used gene expression data from mouse to understand human phenotypes. We believe our approach is appropriate for several reasons. (A) Crucially, the key findings replicated in human data. (B) Single-cell RNA-seq is achievable in mouse but difficult in human neurons (where single-nuclei RNA-seq is typical 42,45,51,52). In brain, differences between single-cell and single-nuclei RNA-seq are important as transcripts that are missed by sequencing nuclei are important for psychiatric disorders, and we previously showed that dendritically-transported transcripts (important for schizophrenia) are specifically depleted from nuclei datasets 12 (we confirmed this finding in four additional datasets, Figure S19). (C) Correlations in gene expression for cell type across species is high (median correlation 0.68, Figure S21), and as high or higher than correlations across methods within cell type and species (single-cell vs single-nuclei RNA-seq, median correlation 0.6) 53. (D) We evaluated protein-coding genes with 1:1 orthologs between mouse and
