RESUMEN
Stuttering is a neuro-developmental speech impairment characterized by uncontrolled utterances (interjections) and core behaviors (blocks, repetitions, and prolongations), and is caused by the failure of speech sensorimotors. Due to its complex nature, stuttering detection (SD) is a difficult task. If detected at an early stage, it could facilitate speech therapists to observe and rectify the speech patterns of persons who stutter (PWS). The stuttered speech of PWS is usually available in limited amounts and is highly imbalanced. To this end, we address the class imbalance problem in the SD domain via a multi-branching (MB) scheme and by weighting the contribution of classes in the overall loss function, resulting in a huge improvement in stuttering classes on the SEP-28 k dataset over the baseline (StutterNet). To tackle data scarcity, we investigate the effectiveness of data augmentation on top of a multi-branched training scheme. The augmented training outperforms the MB StutterNet (clean) by a relative margin of 4.18% in macro F1-score ( F1). In addition, we propose a multi-contextual (MC) StutterNet, which exploits different contexts of the stuttered speech, resulting in an overall improvement of 4.48% in F1 over the single context based MB StutterNet. Finally, we have shown that applying data augmentation in the cross-corpora scenario can improve the overall SD performance by a relative margin of 13.23% in F1 over the clean training.
Asunto(s)
Aprendizaje Profundo , Tartamudeo , Humanos , Tartamudeo/diagnóstico , HablaRESUMEN
Cone photoreceptor dysfunction represents a clinically heterogenous group of disorders characterized by nystagmus, photophobia, reduced central or color vision, and macular dystrophy. Here, we described the molecular findings and clinical manifestations of achromatopsia, a partial or total absence of color vision, co-segregating with three known missense variants of CNGA3 in three large consanguineous Pakistani families. Fundus examination and optical coherence tomography (OCT) imaging revealed myopia, thin retina, retinal pigment epithelial cells loss at fovea/perifovea, and macular atrophy. Combination of Sanger and whole exome sequencing revealed three known homozygous missense variants (c.827A>G, p.(Asn276Ser); c.847C>T, p.(Arg283Trp); c.1279C>T, p.(Arg427Cys)) in CNGA3, the α-subunit of the cyclic nucleotide-gated cation channel in cone photoreceptor cells. All three variants are predicted to replace evolutionary conserved amino acids, and to be pathogenic by specific in silico programs, consistent with the observed altered membrane targeting of CNGA3 in heterologous cells. Insights from our study will facilitate counseling regarding the molecular and phenotypic landscape of CNGA3-related cone dystrophies.
Asunto(s)
Defectos de la Visión Cromática , Células Fotorreceptoras Retinianas Conos , Defectos de la Visión Cromática/diagnóstico , Defectos de la Visión Cromática/genética , Defectos de la Visión Cromática/metabolismo , Canales Catiónicos Regulados por Nucleótidos Cíclicos/genética , Humanos , Mutación , PakistánRESUMEN
Purpose: Cone rod dystrophy (CRD) is a group of inherited retinopathies characterized by the loss of cone and rod photoreceptor cells, which results in poor vision. This study aims to clinically and genetically characterize the segregating CRD phenotype in two large, consanguineous Pakistani families. Methods: Funduscopy, optical coherence tomography (OCT), electroretinography (ERG), color vision, and visual acuity assessments were performed to evaluate the retinal structure and function of the affected individuals. Exome sequencing was performed to identify the genetic cause of CRD. Furthermore, the mutation's effect was evaluated using purified, bacterially expressed ADP-ribosylation factor-like protein 3 (ARL3) and mammalian cells. Results: Fundus photography and OCT imaging demonstrated features that were consistent with CRD, including bull's eye macular lesions, macular atrophy, and central photoreceptor thinning. ERG analysis demonstrated moderate to severe reduction primarily of photopic responses in all affected individuals, and scotopic responses show reduction in two affected individuals. The exome sequencing revealed a novel homozygous variant (c.296G>T) in ARL3, which is predicted to substitute an evolutionarily conserved arginine with isoleucine within the encoded protein GTP-binding domain (R99I). The functional studies on the bacterial and heterologous mammalian cells revealed that the arginine at position 99 is essential for the stability of ARL3. Conclusions: Our study uncovers an additional CRD gene and assigns the CRD phenotype to a variant of ARL3. The results imply that cargo transportation in photoreceptors as mediated by the ARL3 pathway is essential for cone and rod cell survival and vision in humans.
Asunto(s)
Factores de Ribosilacion-ADP/genética , Distrofias de Conos y Bastones/genética , Genes Recesivos , Mutación Puntual , Adolescente , Adulto , Animales , Células COS , Niño , Chlorocebus aethiops , Distrofias de Conos y Bastones/patología , Consanguinidad , Electrorretinografía , Femenino , Expresión Génica , Células HeLa , Homocigoto , Humanos , Masculino , Oftalmoscopía , Linaje , Fenotipo , Tomografía de Coherencia Óptica , Agudeza Visual , Secuenciación del Exoma , Adulto JovenRESUMEN
Purpose: Primary congenital glaucoma (PCG) is a clinically and genetically heterogeneous disease. The present study was undertaken to find the genetic causes of PCG segregating in 36 large consanguineous Pakistani families. Methods: Ophthalmic examination including fundoscopy, or slit-lamp microscopy was performed to clinically characterize the PCG phenotype. Genomic nucleotide sequences of the CYP1B1 and LTBP2 genes were analyzed with either Sanger or whole exome sequencing. In silico prediction programs were used to assess the pathogenicity of identified alleles. ClustalW alignments were performed to determine evolutionary conservation, and three-dimensional (3D) modeling was performed using HOPE and Phyre2 software. Results: Among the known loci, mutations in CYP1B1 and LTBP2 are the common causes of PCG. Therefore, we analyzed the genomic nucleotide sequences of CYP1B1 and LTBP2, and detected probable pathogenic variants cosegregating with PCG in 14 families. These included the three novel (c.542T>A, c.1436A>G, and c.1325delC) and five known (c.868dupC, c.1168C>T, c.1169G>A, c.1209InsTCATGCCACC, and c.1310C>T) variants in CYP1B1. Two of the novel variants are missense substitutions [p.(Leu181Gln), p.(Gln479Arg)], which replaced evolutionary conserved amino acids, and are predicted to be pathogenic by various in silico programs, while the third variant (c.1325delC) is predicted to cause reading frameshift and premature truncation of the protein. A single mutation, p.(Arg390His), causes PCG in six (~43%) of the 14 CYP1B1 mutations harboring families, and thus, is the most common variant in this cohort. Surprisingly, we did not find any LTBP2 pathogenic variants in the families, which further supports the genetic heterogeneity of PCG in the Pakistani population. Conclusions: In conclusion, results of the present study enhance our understanding of the genetic basis of PCG, support the notion of a genetic modifier of CYP1B1, and contribute to the development of genetic testing protocols and genetic counseling for PCG in Pakistani families.