GestaltMML: Enhancing Rare Genetic Disease Diagnosis through Multimodal Machine Learning Combining Facial Images and Clinical Texts.

Individuals with suspected rare genetic disorders often undergo multiple clinical evaluations, imaging studies, laboratory tests and genetic tests, to find a possible answer over a prolonged period of time. Addressing this "diagnostic odyssey" thus has substantial clinical, psychosocial, and economic benefits. Many rare genetic diseases have distinctive facial features, which can be used by artificial intelligence algorithms to facilitate clinical diagnosis, in prioritizing candidate diseases to be further examined by lab tests or genetic assays, or in helping the phenotype-driven reinterpretation of genome/exome sequencing data. Existing methods using frontal facial photos were built on conventional Convolutional Neural Networks (CNNs), rely exclusively on facial images, and cannot capture non-facial phenotypic traits and demographic information essential for guiding accurate diagnoses. Here we introduce GestaltMML, a multimodal machine learning (MML) approach solely based on the Transformer architecture. It integrates facial images, demographic information (age, sex, ethnicity), and clinical notes (optionally, a list of Human Phenotype Ontology terms) to improve prediction accuracy. Furthermore, we also evaluated GestaltMML on a diverse range of datasets, including 528 diseases from the GestaltMatcher Database, several in-house datasets of Beckwith-Wiedemann syndrome (BWS, over-growth syndrome with distinct facial features), Sotos syndrome (overgrowth syndrome with overlapping features with BWS), NAA10-related neurodevelopmental syndrome, Cornelia de Lange syndrome (multiple malformation syndrome), and KBG syndrome (multiple malformation syndrome). Our results suggest that GestaltMML effectively incorporates multiple modalities of data, greatly narrowing candidate genetic diagnoses of rare diseases and may facilitate the reinterpretation of genome/exome sequencing data.

Genomic Contributors to Esophageal Atresia and Tracheoesophageal Fistula: A 12 Year Retrospective Review.

OBJECTIVE: To evaluate genetic testing utilization and diagnostic yield in infants with esophageal atresia (EA)/tracheoesophageal fistula (TEF) over the past 12 years to inform future practices and individualize prognostication and management. STUDY DESIGN: A retrospective cohort study was performed for all infants with EA or EA/TEF hospitalized between January 2011 and January 2023 at a quaternary children's hospital. For each infant, demographic information, prenatal and postnatal history, and genetic testing were reviewed. RESULTS: There were 212 infants who were classified as follows: 1) complex/syndromic with EA/TEF plus an additional major anatomic anomaly (n = 114, of which 74 met VACTERL criteria); 2) isolated/nonsyndromic EA/TEF (n = 88) and 3) isolated/nonsyndromic EA (n = 10). A range of genetic tests were sent with varying diagnostic rates including karyotype analysis in 12 (all with complex/syndromic phenotypes and all positive), chromosomal microarray analysis in 189 (114 of whom were complex/syndromic with an overall diagnostic rate of 3/189), single gene testing for CHD7 in 18 (4 positive), and exome analysis in 37 complex/syndromic patients (8 positive). CONCLUSIONS: EA/TEF with and without additional anomalies is genetically heterogeneous with a broad range of associated phenotypes. While the genetic etiology of EA/TEF with or without VACTERL remains largely unknown, genome wide testing (exome or genome) including copy number analysis is recommended over chromosomal microarray testing. We anticipate that expanded genetic/genomic testing modalities such as RNA sequencing and tissue specific molecular testing are needed in this cohort to improve our understanding of the genomic contributors to EA/TEF.

Recurrent missense variant identified in two unrelated families with MPZL2-related hearing loss, expanding the variant spectrum associated with DFNB111.

MPZL2-related hearing loss is a rare form of autosomal recessive hearing loss characterized by progressive, mild sloping to severe sensorineural hearing loss. Thirty-five previously reported patients had biallelic truncating variants in MPZL2, with the exception of one patient with a missense variant of uncertain significance and a truncating variant. Here, we describe the clinical characteristics and genotypes of five patients from four families with confirmed MPZL2-related hearing loss. A rare missense likely pathogenic variant [NM_005797.4(MPZL2):c.280C>T,p.(Arg94Trp)] located in exon 3 was confirmed to be in trans with a recurrent pathogenic truncating variant that segregated with hearing loss in three of the patients from two unrelated families. This is the first recurrent likely pathogenic missense variant identified in MPZL2. Apparently milder or later-onset hearing loss associated with rare missense variants in MPZL2 indicates that some missense variants in this gene may cause a milder phenotype than that resulting from homozygous or compound heterozygous truncating variants. This study, along with the identification of truncating loss of function and missense MPZL2 variants in several diverse populations, suggests that MPZL2-related hearing loss may be more common than previously appreciated and demonstrates the need for MPZL2 inclusion in hearing loss testing panels.

Genomic analyses in Cornelia de Lange Syndrome and related diagnoses: Novel candidate genes, genotype-phenotype correlations and common mechanisms.

Cornelia de Lange Syndrome (CdLS) is a rare, dominantly inherited multisystem developmental disorder characterized by highly variable manifestations of growth and developmental delays, upper limb involvement, hypertrichosis, cardiac, gastrointestinal, craniofacial, and other systemic features. Pathogenic variants in genes encoding cohesin complex structural subunits and regulatory proteins (NIPBL, SMC1A, SMC3, HDAC8, and RAD21) are the major pathogenic contributors to CdLS. Heterozygous or hemizygous variants in the genes encoding these five proteins have been found to be contributory to CdLS, with variants in NIPBL accounting for the majority (>60%) of cases, and the only gene identified to date that results in the severe or classic form of CdLS when mutated. Pathogenic variants in cohesin genes other than NIPBL tend to result in a less severe phenotype. Causative variants in additional genes, such as ANKRD11, EP300, AFF4, TAF1, and BRD4, can cause a CdLS-like phenotype. The common role that these genes, and others, play as critical regulators of developmental transcriptional control has led to the conditions they cause being referred to as disorders of transcriptional regulation (or "DTRs"). Here, we report the results of a comprehensive molecular analysis in a cohort of 716 probands with typical and atypical CdLS in order to delineate the genetic contribution of causative variants in cohesin complex genes as well as novel candidate genes, genotype-phenotype correlations, and the utility of genome sequencing in understanding the mutational landscape in this population.

Stimulating Duddingtonia flagrans chlamydospore production through dehydration.

Duddingtonia flagrans is a nematode-trapping fungus that has shown promising results as a tool to combat parasitic nematode infections in livestock. The fungus interrupts the parasitic lifecycle by trapping and killing larval stages on pasture to prevent re-infection of animals. One barrier to the fungus' commercial use is scaling up production of the fungus, and specifically of chlamydospores, which survive the digestive tract to grow in fecal pats on pasture, thus have potential as a feed through anthelmintic. The purpose of this study was to evaluate the effect of dehydration on sporulation of the fungus. Disks of Duddingtonia flagrans type strain (ATCC® 13423™) were grown on 17% cornmeal agar for 26 days at 30 °C, then split into three groups; dried quickly at 38 °C and 37% humidity over 48 h ("incubated"), dried more slowly at 24 °C and 55% humidity over 10 days ("air-dried"), or kept at 30 °C and sealed with parafilm to prevent loss of moisture as a control ("wet"). Half of each dried culture was resuspended in water, then heated to liquify and homogenized through vortexing. Spores were then counted in a Neubauer hematocytometer. Both the "air-dried" and "incubated" drying techniques yielded significantly more spores than the "wet" control (Welch's two sample t test p values of .0359 and .0411, respectively). The difference in average chlamydospores per milliliter was insignificant between the two drying techniques, although a visual representation of the data shows less spore count variability in the "air-dried" technique.

Effects of smoking abstinence on smoking-reinforced responding, withdrawal, and cognition in adults with and without attention deficit hyperactivity disorder.

RATIONALE: Individuals with attention deficit hyperactivity disorder (ADHD) have a more difficult time quitting smoking compared to their non-ADHD peers. Little is known about the underlying behavioral mechanisms associated with this increased risk. OBJECTIVES: This study aims to assess the effects of 24-h smoking abstinence in adult smokers with and without ADHD on the following outcomes: smoking-reinforced responding, withdrawal, and cognitive function. METHODS: Thirty-three (n = 16 with ADHD, 17 without ADHD) adult smokers (more than or equal to ten cigarettes/day) were enrolled. Each participant completed two experimental sessions: one following smoking as usual and one following biochemically verified 24-h smoking abstinence. Smoking-reinforced responding measured via a progressive ratio task, smoking withdrawal measured via questionnaire, and cognition measured via a continuous performance test (CPT) were assessed at each session. RESULTS: Smoking abstinence robustly increased responding for cigarette puffs in both groups, and ADHD smokers responded more for puffs regardless of condition. Males in both groups worked more for cigarette puffs and made more commission errors on the CPT than females, regardless of condition. Smoking abstinence also increased ratings of withdrawal symptoms in both groups and smokers with ADHD, regardless of condition, reported greater symptoms of arousal, habit withdrawal, and somatic complaints. Across groups, smoking abstinence decreased inhibitory control and increased reaction time variability on the CPT. Abstinence-induced changes in inhibitory control and negative affect significantly predicted smoking-reinforced responding across groups. CONCLUSIONS: Smokers with ADHD reported higher levels of withdrawal symptoms and worked more for cigarette puffs, regardless of condition, which could help explain higher levels of nicotine dependence and poorer cessation outcomes in this population. Abstinence-induced changes in smoking-reinforced responding are associated with changes in inhibitory control and negative affect regardless of ADHD status, a finding that may lead to novel prevention and treatment programs.