Answers are grounded in this paper and cite their sources.
Autism spectrum disorder (ASD) is a heterogeneous neurodevelopmental condition shaped by de novo protein‐truncating variants, inherited missense substitutions, recurrent copy number variations (CNVs), and distributed polygenic risk. These alterations converge on a limited set of neurobiological processes, including synaptic refinement, excitation‐to‐inhibition balance, chromatin state regulation, and glial‐mediated neuroimmune signaling. Experimental paradigms now exist for systematic variant‐to‐function interrogation through CRISPR‐engineered isogenic human pluripotent stem‐cell libraries incorporating Cas9 nucleases, programmable transcriptional modulators (dCas9‐KRAB and dCas9‐VP64), base editors, and prime editing systems. Coupled with single‐cell multi‐omic platforms such as Perturb‐seq, scATAC‐seq, and CITE‐seq, and with circuit‐level phenotyping through multielectrode arrays, calcium imaging, and quantitative morphometry, these systems resolve ASD‐relevant cellular states at single‐cell and cell‐type‐specific resolution. In this narrative review, we distinguish deliberately between what these platforms have already established and what remains proposed. To date, a small number of studies constitute the demonstrated evidence base: in vivo Perturb‐seq of 35 ASD/neurodevelopmental risk genes in mouse cortex, mosaic knockout of ASD genes in human cortical organoids, the CHOOSE organoid screen of 36 risk genes, assembloid CRISPR screening of 425 neurodevelopmental‐disorder risk genes, a 63‐line mouse embryonic stem‐cell CNV resource, and a CRISPRi screen of 102 ASD risk genes in human induced pluripotent stem‐cell (iPSC)‐derived microglia implicating ADNP in synaptic pruning. We summarized these in a dedicated comparison table. Against this, integration of polygenic and noncoding architecture, causal‐inference frameworks, such as Mendelian randomization and do‐calculus, and cross‐laboratory reproducibility, remain aspirational rather than realized. We therefore frame the field's principal near‐term requirement not as further gene discovery but as harmonized protocols, shared reference lines, arrayed validation capacity, and interpretable statistical frameworks capable of converting perturbation atlases into causal mechanism.
No takes yet. Share an insight, caveat, or question.
Uddin et al. (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: