Quick appraisal

This preprint presents a pooled CRISPRi NPC proliferation/survival screen that prioritizes 145 genes (24 ASD, 77 CHD, 44 ASD–CHD) and identifies a cilia-enriched cluster; follow‑up experiments validate primary‑cilia and TAOK1 motile‑cilia/heart/brain phenotypes. The screen design, orthogonal in vitro cilia assays, and Xenopus in vivo validations are coherent and mostly well described, but key limitations (preprint status, CRISPRi knockdown levels/dosage effects, morphological quantification thresholds, and reliance on morpholinos) temper claims about human disease causality and clinical stratification. Teerikorpi et al. (preprint)

Visual paper review — “Ciliary biology intersects autism and congenital heart disease” (bioRxiv 2024)

Visualize first — key numerical summary from the paper (raw counts come from the authors' reported screen summaries and tables).

Data source: pooled CRISPRi screen reported in the preprint (361 genes targeted; 145 significant gene-level hits; 24 ASD-only, 77 CHD-only, 44 shared ASD–CHD) Teerikorpi et al., preprint

Visual evidence map — key claims, supporting experiments, and strength

The chain of evidence in the paper runs: pooled genetic perturbation → statistical hit calling → clustering + STRING/ToppGene enrichment → focused CRISPRi imaging of cilia in two cell types → Xenopus TAOK1 morpholino work. Strength of each node varies and is discussed below with citations.

Detailed critique (visual-first, concise bullets, then evidence)

• Screen design & sample handling: good scale (361 genes, 5 guides/gene, 3 replicates, timepoints day 0/5/10/20) but sensitivity to guide efficacy and CRISPRi repression magnitude is an important confounder that the authors partially address with qPCR for follow-ups (Teerikorpi et al., methods).
• Statistical thresholds: FDR < 0.1 is standard for multiplex screens but relatively permissive; authors prioritize Log2FC ≥ 0.585 for further analyses which is transparent and reasonable given the pooled format (Thresholds used in the screen).
• Biological plausibility — cilia link: independent literature strongly connects cilia/microtubules to both neurodevelopment and CHD; the authors' GO enrichment and STRING clustering support this intersection (Primary cilia review;Cilia and CHD (Li et al. 2015)).
• Experimental follow-up — primary cilia assays: authors tested seven ASD–CHD genes (4 with prior cilia links, 3 novel candidates) in RPE1 and NPCs and report decreases in percent ciliated cells and cilia length; methodology (ARL13B staining, CiliaQ, CellProfiler) is appropriate but image/quantification automation details and potential off-target or proliferation-linked secondary effects require careful interpretation (Cilia imaging and quantification).
• In vivo TAOK1 evidence: Xenopus morpholino knockdown produced robust motile cilia loss and reduced telencephalon and heart ventricle size—supporting developmental roles; morpholino caveats (off-targets, compensation, dose sensitivity) mean rescue experiments or CRISPR-based knockout/knock-in would strengthen causality (TAOK1 in Xenopus; patient data mention).
• Translational claim caution: authors propose using CHD diagnoses to stratify ASD risk — biologically plausible but requires population-level longitudinal genetics/phenotyping and effect-size estimates, which are not delivered here; current evidence prioritizes gene‑level candidates rather than clinical predictive models (Translational claim in Discussion).

Methodological blindspots & limitations (concise)

1. CRISPRi repression strength varies by guide/promoter context; pooled screens can miss dosage‑sensitive or haploinsufficient phenotypes and can confound cilia defects with proliferation changes — authors note this and suggest future lower‑dose assays in post-mitotic cells Authors.
2. Morpholino use in Xenopus lacks orthogonal CRISPR genetic validation and rescue experiments shown explicitly; morpholino phenotypes can be confounded by toxicity/off-target effects — standard control rescue (human mRNA) would mitigate this concern.
3. No population penetrance/segregation/variant-level statistics — the screen prioritizes genes but not which classes of human variants (LoF vs missense, de novo vs inherited) confer joint ASD+CHD risk; concluding shared risk for the gene is plausible but not proven for human patients without genotype-phenotype data integration.
4. Potential selection biases in gene lists: ASD and CHD gene lists derive from specific prior exome studies and network predictions — biases in those sources propagate into which genes were screened and prioritized (authors use Satterstrom 2020, Jin 2017 lists and Rosenthal network predictions) Satterstrom et al. (ASD gene list)).

What would change the main conclusion?

• Independent genetic evidence showing that deleterious (LoF) variants in the prioritized CHD genes increase ASD prevalence in large cohorts (case-control or family-based) would strengthen the shared‑risk claim.
• CRISPR/Cas9 knockout (and rescue) experiments in Xenopus or mouse that replicate TAOK1 morpholino phenotypes and show dose-dependent penetrance would rule out morpholino artefacts.
• Single-cell post-mitotic neuronal assays showing direct cilia functional deficits leading to altered SHH or cilia-dependent signaling in neural differentiation would provide mechanistic depth linking cilia defects to neurodevelopmental outcomes.

Practical, prioritized next experiments (visual, short)

1. Rescue TAOK1 morpholino by co-injecting human TAOK1 mRNA (WT and kinase-dead) — distinguishes specificity and kinase dependence.
2. CRISPR KO/knock-in (patient variants) of TAOK1 in Xenopus and/or zebrafish with quantitative cilia & heart metrics to confirm phenotypes across models.
3. Single-cell transcriptomics (NPC→neurons) with and without TAOK1/CEP290 perturbation to track SHH and cilia-regulated signaling changes during differentiation.
4. Population genetics: analyze large ASD and CHD exome cohorts for enrichment of LoF variants in the 45 CHD genes the screen prioritized.

Bottom-line evaluation (concise)

The paper provides convincing experimental evidence that ciliary biology intersects a subset of genes implicated in ASD and CHD, and it identifies TAOK1 as a compelling shared candidate with in vivo phenotypes. The translational claim (stratifying CHD patients for ASD risk) is promising but premature; further genetic, rescue, and dosage‑sensitive validations are needed to move from candidate prioritization to clinical application Paper.

Author-review quick links

Note: This review relies only on the preprint text, figures and tables; the paper has not (yet) been peer reviewed — treat conclusions as provisional.