Identifying facial phenotypes of genetic disorders using deep learning

Gurovich, Yaron; Hanani, Yair; Bar, Omri; Nadav, Guy; Fleischer, Nicole; Gelbman, Dekel; Basel‐Salmon, Lina; Krawitz, Peter; Kamphausen, Susanne; Zenker, Martin; Bird, Lynne M.; Gripp, Karen W.

doi:10.1038/s41591-018-0279-0

Cited by 472 publications

(402 citation statements)

References 27 publications

Supporting

Mentioning

389

Contrasting

Unclassified

Order By: Relevance

“…Facial features associated with each gene in NS are challenging to describe and both typical and atypical faces were found in each gene category . Recently, artificial intelligence (AI) has been used to predict the genotype of NS using facial photos, with accuracy of 64%, but the experiment was an evenly sized 5 class problem . The average random accuracy of predicting the genotypes of NS in our study is lower than the experiment.…”

Section: Discussionmentioning

confidence: 60%

Molecular and phenotypic spectrum of Noonan syndrome in Chinese patients

Yao

Tan

et al. 2019

Clinical Genetics

View full text Add to dashboard Cite

Noonan syndrome (NS) is a common autosomal dominant/recessive disorder. No large‐scale study has been conducted on NS in China, which is the most populous country in the world. Next‐generation sequencing (NGS) was used to identify pathogenic variants in patients that exhibited NS‐related phenotypes. We assessed the facial features and clinical manifestations of patients with pathogenic or likely pathogenic variants in the RAS‐MAPK signaling pathway. Gene‐related Chinese NS facial features were described using artificial intelligence (AI).NGS identified pathogenic variants in 103 Chinese patients in eight NS‐related genes: PTPN11 (48.5%), SOS1 (12.6%), SHOC2 (11.7%), KRAS (9.71%), RAF1 (7.77%), RIT1 (6.8%), CBL (0.97%), NRAS (0.97%), and LZTR1 (0.97%). Gene‐related facial representations showed that each gene was associated with different facial details. Eight novel pathogenic variants were detected and clinical features because of specific genetic variants were reported, including hearing loss, cancer risk due to a PTPN11 pathogenic variant, and ubiquitous abnormal intracranial structure due to SHOC2 pathogenic variants. NGS facilitates the diagnosis of NS, especially for patients with mild/moderate and atypical symptoms. Our study describes the genotypic and phenotypic spectra of NS in China, providing new insights into distinctive clinical features due to specific pathogenic variants.

show abstract

Section: Discussionmentioning

confidence: 60%

Molecular and phenotypic spectrum of Noonan syndrome in Chinese patients

Yao

Tan

et al. 2019

Clinical Genetics

View full text Add to dashboard Cite

show abstract

“…The clinical diagnosis of KS may be facilitated through deep phenotyping using HPO terminology (Groza et al, ), the use of diagnostic clinical criteria (Adam et al, ; Makrythanasis et al, ) and/or the use of artificial intelligence and deep learning algorithms in the setting of facial recognition (Gurovich et al, ). The latter was found to be useful in identifying the correct diagnosis retrospectively using a facial image and clinical phenotype information.…”

Section: Discussionmentioning

confidence: 99%

“…Although its phenotypic spectrum is relatively well characterized in Caucasian and Japanese populations, only limited and incomplete information is available in Chinese populations (Guo et al, 2018). In the present study we report the detailed phenotypes of 14 patients with KS from two tertiary The clinical diagnosis of KS may be facilitated through deep phenotyping using HPO terminology (Groza et al, 2015), the use of diagnostic clinical criteria (Adam et al, 2019;Makrythanasis et al, 2013) and/or the use of artificial intelligence and deep learning algorithms in the setting of facial recognition (Gurovich et al, 2019). The latter was found to be useful in identifying the correct diagnosis retrospectively using a facial image and clinical phenotype information.…”

Section: Discussionmentioning

confidence: 99%

The phenotypic spectrum of Kabuki syndrome in patients of Chinese descent: A case series

Wang

Niu

et al. 2019

American J of Med Genetics Pt A

View full text Add to dashboard Cite

Kabuki syndrome (KS) is a rare disorder of transcriptional regulation with a complex phenotype that includes cranio‐facial dysmorphism, intellectual disability, hypotonia, failure to thrive, short stature, and cardiac and renal anomalies. Heterozygous, de novo dominant mutations in either KMT2D or KDM6A underlie KS. Limited information is available about the phenotypic spectrum of KS in China. Fourteen Chinese patients with genetically confirmed KS were evaluated in addition to 11 Chinese patients who were identified from the medical literature. The clinical phenotype spectrum of these patients was compared to that of 449 patients with KS from non‐Chinese ethnicities. Additionally, we explored the utility of a facial recognition software in recognizing KS. All 25 patients with KS carried de novo, likely pathogenic or pathogenic variants in either KMT2D or KDM6A. Eighteen patients were male, the age at diagnosis ranged from 2months to 11.6 years. The facial gestalt included arched and broad eyebrows (25/25; 100%), sparse lateral or notched eyebrows (18/18; 100%), short columella with a concave nasal tip (24/25; 96%) and large, prominent ears (24/24; 100%) which were more frequent in Chinese patients (p < .01). In contrast, microcephaly (2/25; 8%), cleft lip/palate (2/25; 8%), and cardiac defects (10/25; 40%) were less frequent in Chinese patients (not significant). The diagnosis of KS was correctly identified in 13 of 14 patients through facial recognition and clinical phenotyping, underscoring the utility of this approach. As expected, there is marked phenotypic overlap between Chinese and non‐Chinese patients with KS, although subtle differences were identified.

show abstract

“…Both probands showed a reduced expression of total NBAS, compatible with NMD, associated Nonspecific dysmorphic features have been previously reported in patients with biallelic pathogenic NBAS variants(Staufner et al, 2016). To explore this issue systematically, we compared 28 frontal facial images collected from 16 patients (published and present cases; cohort including 10 females and 6 males, age range: 0.6-37 years, median age: 9 years, ethnicity: 14 Europeans and 2 Arabs) and unaffected controls using the Face2Gene research application(version 18.1.8; www.face2gene.com), with the facial recognition technology called DeepGestalt (FDNA Inc, Boston, MA), which has been described byGurovich et al (2019). Controls included a total of 16 unrelated age-and gender-matched individuals.…”

mentioning

confidence: 99%

NBAS pathogenic variants: Defining the associated clinical and facial phenotype and genotype–phenotype correlations

Carli¹,

Giorgio

Pantaleoni

et al. 2019

Human Mutation

Self Cite

View full text Add to dashboard Cite

The pathogenic variants in the neuroblastoma‐amplified sequence (NBAS) are associated with a clinical spectrum involving the hepatic, skeletal, ocular, and immune systems. Here, we report on two unrelated subjects with a complex phenotype solved by whole‐exome sequencing, who shared a synonymous change in NBAS that was documented to affect the transcript processing and co‐occurring with a truncating change. Starting from these two cases, we systematically assessed the clinical information available for all subjects with biallelic NBAS pathogenic variants (73 cases in total). We revealed a recognizable facial profile (hypotelorism, thin lips, pointed chin, and “progeroid” appearance) determined by using DeepGestalt facial recognition technology, and we provide evidence for the occurrence of genotype–phenotype correlations. Notably, severe hepatic involvement was associated with variants affecting the NBAS‐Nter and Sec39 domains, whereas milder liver involvement and immunodeficiency were generally associated with variants located at the N‐terminus and C‐terminus of the protein. Remarkably, no patient was reported to carry two nonsense variants, suggesting lethality of complete NBAS loss‐of‐function.

show abstract

Identifying facial phenotypes of genetic disorders using deep learning

Cited by 472 publications

References 27 publications

Molecular and phenotypic spectrum of Noonan syndrome in Chinese patients

Molecular and phenotypic spectrum of Noonan syndrome in Chinese patients

The phenotypic spectrum of Kabuki syndrome in patients of Chinese descent: A case series

NBAS pathogenic variants: Defining the associated clinical and facial phenotype and genotype–phenotype correlations

Contact Info

Product

Resources

About