Selected Grantee Publications
Deep Learning Is Widely Applicable to Phenotyping Embryonic Development and Disease
Naert et al., Development. 2021.
https://pubmed.ncbi.nlm.nih.gov/34739029/
Genome editing simplifies the generation of new animal models for congenital disorders. The authors illustrate how deep learning (U-Net) automates segmentation tasks in various imaging modalities. They demonstrate this approach in embryos with polycystic kidneys (pkd1 and pkd2) and craniofacial dysmorphia (six1). They provide a library of pre-trained networks and detailed instructions for applying deep learning to datasets and demonstrate the versatility, precision, and scalability of deep neural network phenotyping on embryonic disease models. Supported by ORIP (P40OD010997, R24OD030008), NICHD, NIDDK, and NIMH.
Sequence Diversity Analyses of an Improved Rhesus Macaque Genome Enhance its Biomedical Utility
Warren et al., Science. 2020.
https://science.sciencemag.org/content/370/6523/eabc6617
Investigators sequenced and assembled an Indian-origin female rhesus macaque (RM) genome using a multiplatform genomics approach that included long-read sequencing, extensive manual curation, and experimental validation to generate a new comprehensive annotated reference genome. As a result, 99.7% of the gaps in the earlier draft genome are now closed, and more than 99% of the genes are represented. Whole-genome sequencing of 853 RMs of both sexes identified 85.7 million single-nucleotide variants and 10.5 million indel variants, including potentially damaging variants in genes associated with human autism and developmental delay. The improved assembly of segmental duplications, new lineage-specific genes and expanded gene families provide a framework for developing noninvasive NHP models for human disease, as well as studies of genetic variation and phenotypic consequences. Supported by ORIP (P51OD011106, P51OD011107, P51OD011132, P51OD011104, U42OD024282, U42OD010568, R24OD011173, R24OD021324, R24OD010962), NHGRI, NIMH, NHLBI, and NIGMS.