Editors' ChoiceAutism

Right Place, Right Time

See allHide authors and affiliations

Science Translational Medicine  04 Dec 2013:
Vol. 5, Issue 214, pp. 214ec202
DOI: 10.1126/scitranslmed.3008007

When new loci are implicated in human disease, further inquiry is needed to translate these findings into bench and bedside applications: What are the causal genes within the identified loci, and in what networks, tissues, cell types, and time periods do they act to contribute to disease? Answering these questions can be challenging, but the work of Willsey et al. on autism demonstrates that success can be achieved through collection and systematic analysis of multiple large-scale data sets.

The researchers seeded their study with nine genes implicated in autism, on the basis of at least two loss-of-function (LOF) de novo mutations identified from autism probands and coexpression data from the BrainSpan project, which spans 15 consecutive periods of neurodevelopment from fetal stages into adulthood. Integration and enrichment analysis of these data jointly identified three spatiotemporal networks associated with autism. The researchers then found that additional de novo LoF mutations in affected probands were enriched in other genes found within the three implicated networks. Next, they queried their networks using human brain layer–specific expression data from BrainSpan and coexpression assays in mouse brain in order to identify the specific layer of brain (the inner cortical plate) and the time period (midfetal,10 to 19 weeks after conception) important for pathogenicity. Last, they used a set of marker genes to distinguish among multiple cell types spanning superficial and deep cortical layers of the developing human brain, finding enrichment in cortical glutamatergic projection neurons, specific to the deep cortical layer.

The researchers suggest that additional gene networks will likely be implicated in risk to autism, but this work of Willsey et al. is an important initial advance, identifying genes acting in the right place and the right time for autism predisposition. That these important findings were initiated through careful integration and analysis of multiple large public data sets highlights the promise of big data and the need for the development of tools that enable researchers to meaningfully integrate and interrogate these data.

A. J. Willsey et al., Coexpression networks implicate human midfetal deep cortical projection neurons in the pathogenesis of autism. Cell 155, 997–1007 (2013). [Abstract]

Navigate This Article