RESUMEN
Predicting absorption, distribution, metabolism, excretion, and toxicity (ADMET) properties of small molecules is a key task in drug discovery. A major challenge in building better ADMET models is the experimental error inherent in the data. Furthermore, ADMET predictors are typically regression tasks due to the continuous nature of the data, which makes it difficult to apply existing denoising methods from other domains as they largely focus on classification tasks. Here, we develop denoising schemes based on deep learning to address this. We find that the training error (TE) can be used to identify the noise in regression tasks while ensemble-based and forgotten event-based metrics fail to detect the noise. The most significant performance increase occurs when the original model is finetuned with the denoised data using TE as the noise detection metric. Our method has the ability to improve models with medium noise and does not degrade the performance of models with noise outside this range (low noise and high noise regimes). To our knowledge, our denoising scheme is the first to improve model performance for ADMET data and has implications for improving models for experimental assay data in general.
RESUMEN
Mammalian inner ear hair cell loss leads to permanent hearing and balance dysfunction. In contrast to the cochlea, vestibular hair cells of the murine utricle have some regenerative capacity. Whether human utricular hair cells regenerate in vivo remains unknown. Here we procured live, mature utricles from organ donors and vestibular schwannoma patients, and present a validated single-cell transcriptomic atlas at unprecedented resolution. We describe markers of 13 sensory and non-sensory cell types, with partial overlap and correlation between transcriptomes of human and mouse hair cells and supporting cells. We further uncover transcriptomes unique to hair cell precursors, which are unexpectedly 14-fold more abundant in vestibular schwannoma utricles, demonstrating the existence of ongoing regeneration in humans. Lastly, supporting cell-to-hair cell trajectory analysis revealed 5 distinct patterns of dynamic gene expression and associated pathways, including Wnt and IGF-1 signaling. Our dataset constitutes a foundational resource, accessible via a web-based interface, serving to advance knowledge of the normal and diseased human inner ear.
Asunto(s)
Regeneración , Análisis de la Célula Individual , Transcriptoma , Humanos , Animales , Regeneración/genética , Ratones , Sáculo y Utrículo/metabolismo , Sáculo y Utrículo/citología , Neuroma Acústico/genética , Neuroma Acústico/metabolismo , Neuroma Acústico/patología , Oído Interno/metabolismo , Oído Interno/citología , Factor I del Crecimiento Similar a la Insulina/metabolismo , Factor I del Crecimiento Similar a la Insulina/genética , Masculino , Células Ciliadas Vestibulares/metabolismo , Femenino , Perfilación de la Expresión GénicaRESUMEN
Antibody thermostability is challenging to predict from sequence and/or structure. This difficulty is likely due to the absence of direct entropic information. Herein, we present AbMelt where we model the inherent flexibility of homologous antibody structures using molecular dynamics simulations at three temperatures and learn the relevant descriptors to predict the temperatures of aggregation (Tagg), melt onset (Tm,on), and melt (Tm). We observed that the radius of gyration deviation of the complementarity determining regions at 400 K is the highest Pearson correlated descriptor with aggregation temperature (rp = -0.68 ± 0.23) and the deviation of internal molecular contacts at 350 K is the highest correlated descriptor with both Tm,on (rp = -0.74 ± 0.04) as well as Tm (rp = -0.69 ± 0.03). Moreover, after descriptor selection and machine learning regression, we predict on a held-out test set containing both internal and public data and achieve robust performance for all endpoints compared with baseline models (Tagg R2 = 0.57 ± 0.11, Tm,on R2 = 0.56 ± 0.01, and Tm R2 = 0.60 ± 0.06). In addition, the robustness of the AbMelt molecular dynamics methodology is demonstrated by only training on <5% of the data and outperforming more traditional machine learning models trained on the entire data set of more than 500 internal antibodies. Users can predict thermostability measurements for antibody variable fragments by collecting descriptors and using AbMelt, which has been made available.
RESUMEN
Pretrained deep learning models self-supervised on large datasets of language, image, and graph representations are often fine-tuned on downstream tasks and have demonstrated remarkable adaptability in a variety of applications including chatbots, autonomous driving, and protein folding. Additional research aims to improve performance on downstream tasks by fusing high dimensional data representations across multiple modalities. In this work, we explore a novel fusion of a pretrained language model, ChemBERTa-2, with graph neural networks for the task of molecular property prediction. We benchmark the MolPROP suite of models on seven scaffold split MoleculeNet datasets and compare with state-of-the-art architectures. We find that (1) multimodal property prediction for small molecules can match or significantly outperform modern architectures on hydration free energy (FreeSolv), experimental water solubility (ESOL), lipophilicity (Lipo), and clinical toxicity tasks (ClinTox), (2) the MolPROP multimodal fusion is predominantly beneficial on regression tasks, (3) the ChemBERTa-2 masked language model pretraining task (MLM) outperformed multitask regression pretraining task (MTR) when fused with graph neural networks for multimodal property prediction, and (4) despite improvements from multimodal fusion on regression tasks MolPROP significantly underperforms on some classification tasks. MolPROP has been made available at https://github.com/merck/MolPROP . SCIENTIFIC CONTRIBUTION: This work explores a novel multimodal fusion of learned language and graph representations of small molecules for the supervised task of molecular property prediction. The MolPROP suite of models demonstrates that language and graph fusion can significantly outperform modern architectures on several regression prediction tasks and also provides the opportunity to explore alternative fusion strategies on classification tasks for multimodal molecular property prediction.
RESUMEN
MOTIVATION: Pre-trained protein language and/or structural models are often fine-tuned on drug development properties (i.e. developability properties) to accelerate drug discovery initiatives. However, these models generally rely on a single structural conformation and/or a single sequence as a molecular representation. We present a physics-based model, whereby 3D conformational ensemble representations are fused by a transformer-based architecture and concatenated to a language representation to predict antibody protein properties. Antibody language ensemble fusion enables the direct infusion of thermodynamic information into latent space and this enhances property prediction by explicitly infusing dynamic molecular behavior that occurs during experimental measurement. RESULTS: We showcase the antibody language ensemble fusion model on two developability properties: hydrophobic interaction chromatography retention time and temperature of aggregation (Tagg). We find that (i) 3D conformational ensembles that are generated from molecular simulation can further improve antibody property prediction for small datasets, (ii) the performance benefit from 3D conformational ensembles matches shallow machine learning methods in the small data regime, and (iii) fine-tuned large protein language models can match smaller antibody-specific language models at predicting antibody properties. AVAILABILITY AND IMPLEMENTATION: AbLEF codebase is available at https://github.com/merck/AbLEF.
Asunto(s)
Termodinámica , Anticuerpos/química , Conformación Proteica , Aprendizaje Automático , Interacciones Hidrofóbicas e Hidrofílicas , Programas Informáticos , Biología Computacional/métodosRESUMEN
Mycotic aneurysms are rare but potentially catastrophic. We report a case of an innominate artery pseudoaneurysm in a 4-year-old patient that caused a tracheoinnominate fistula requiring tracheoplasty with a costal cartilage graft and a homograft iliac artery replacement of the diseased innominate artery, with a successful outcome.
RESUMEN
Obstructed breathing is the most common indication for tonsillectomy in children. Although tonsillectomy is performed frequently worldwide, the surgery is associated with a number of significant complications such as bleeding and respiratory failure. Complication risk depends on a number of complex factors, including indications for surgery, demographics, patient comorbidities, and variations in perioperative techniques. While polysomnography is currently accepted as the gold standard diagnostic tool for obstructive sleep apnea, studies evaluating outcomes following surgery suggest that more research is needed on the identification of more readily available and accurate tools for the diagnosis and follow-up of children with obstructed breathing.
Asunto(s)
Adenoidectomía , Polisomnografía , Apnea Obstructiva del Sueño , Tonsilectomía , Humanos , Apnea Obstructiva del Sueño/cirugía , Apnea Obstructiva del Sueño/diagnóstico , Tonsilectomía/métodos , Tonsilectomía/efectos adversos , Adenoidectomía/métodos , Adenoidectomía/efectos adversos , Niño , Complicaciones Posoperatorias/etiología , Resultado del TratamientoRESUMEN
INTRODUCTION: Identify non-glycemic factors affecting the relationship between fasting plasma glucose (FPG) and glycated hemoglobin (HbA1c), in order to refine diabetes diagnostic criteria. RESEARCH DESIGN AND METHODS: Relationship between FPG-HbA1c was assessed in 12 531 individuals from 2001 to 2018 US National Health and Nutrition Examination Survey. Using a recently described method, FPG and HbA1c were used to calculate apparent glycation ratio (AGR) of red blood cells for different subgroups based on age, race, and gender. RESULTS: At an FPG of 7 mmol/L, black individuals had a higher HbA1c (p<0.001, mean: 50.2 mmol/mol, 95% CI (49.8 to 50.4)) compared with white individuals (47.4 mmol/mol (47.2 to 47.5)). This corresponds to NGSP (National Glycohemoglobin Standardization Program) units of 6.7% and 6.5% for black versus white individuals, respectively. Similarly, individuals under 21 years had lower HbA1c (p<0.001, 47.9 mmol/mol (47.7 to 48.1), 6.5%) compared with those over 50 years (48.3 mmol/mol (48.2 to 48.5), 6.6%). Differences were also observed between women (p<0.001, 49.2 mmol/mol (49.1 to 49.3), 6.7%) and men (47.0 mmol/mol (46.8 to 47.1), 6.5%). Of note, the difference in HbA1c at FPG of 7 mmol/L in black females over 50 and white males under 21 years was 5 mmol/mol (0.46%). AGR differences according to race (p<0.001), age (p<0.001), and gender (p<0.001) explained altered glucose-HbA1c relationship in the analyzed groups. CONCLUSIONS: FPG-HbA1c relationship is affected by non-glycemic factors leading to incorrect diagnosis of diabetes in some individuals and ethnic groups. Assessment of AGR helps understand individual-specific relationship between glucose levels and HbA1c, which has the potential to more accurately diagnose and manage diabetes.
Asunto(s)
Diabetes Mellitus , Etnicidad , Masculino , Femenino , Humanos , Hemoglobina Glucada , Encuestas Nutricionales , Ayuno , Diabetes Mellitus/diagnóstico , Diabetes Mellitus/epidemiología , GlucosaRESUMEN
Background: Time in range (TIR), time in tight range (TITR), and average glucose (AG) are used to adjust glycemic therapies in diabetes. However, TIR/TITR and AG can show a disconnect, which may create management difficulties. We aimed to understand the factors influencing the relationships between these glycemic markers. Materials and Methods: Real-world glucose data were collected from self-identified diabetes type 1 and type 2 diabetes (T1D and T2D) individuals using flash continuous glucose monitoring (FCGM). The effects of glycemic variability, assessed as glucose coefficient of variation (CV), on the relationship between AG and TIR/TITR were investigated together with the best-fit glucose distribution model that addresses these relationships. Results: Of 29,164 FCGM users (16,367 T1D, 11,061 T2D, and 1736 others), 38,259 glucose readings/individual were available. Comparing low and high CV tertiles, TIR at AG of 150 mg/dL varied from 80% ± 5.6% to 62% ± 6.8%, respectively (P < 0.001), while TITR at AG of 130 mg/dL varied from 65% ± 7.5% to 49% ± 7.0%, respectively (P < 0.001). In contrast, higher CV was associated with increased TIR and TITR at AG levels outside the upper limit of these ranges. Gamma distribution was superior to six other models at explaining AG and TIR/TITR interactions and demonstrated nonlinear interplay between these metrics. Conclusions: The gamma model accurately predicts interactions between CGM-derived glycemic metrics and reveals that glycemic variability can significantly influence the relationship between AG and TIR with opposing effects according to AG levels. Our findings potentially help with clinical diabetes management, particularly when AG and TIR appear mismatched.
Asunto(s)
Automonitorización de la Glucosa Sanguínea , Glucemia , Diabetes Mellitus Tipo 1 , Diabetes Mellitus Tipo 2 , Humanos , Glucemia/análisis , Femenino , Masculino , Diabetes Mellitus Tipo 2/sangre , Persona de Mediana Edad , Diabetes Mellitus Tipo 1/sangre , Diabetes Mellitus Tipo 1/tratamiento farmacológico , Adulto , Control Glucémico , Anciano , Hipoglucemiantes/uso terapéuticoRESUMEN
OBJECTIVE: To assess the utility of diffusion tensor imaging of the auditory pathway in children with sensorineural hearing loss (SNHL). STUDY DESIGN: Retrospective cohort study. SETTING: A single academic tertiary children's hospital. PATIENTS: Sixteen pediatric patients with bilateral SNHL of at least moderate severity in the poorer ear (eight male; mean age, 5.3 ± 4.9 yrs). Controls consisted of age- and sex-matched children with normal hearing who were imaged for nonotologic, non-neurologic medical concerns and found to have normal magnetic resonance imaging (MRI). INTERVENTIONS: Three Tesla MRI scanners were used for diffusion tensor imaging. MAIN OUTCOME MEASURES: Quantitative diffusion tensor metrics were extracted from the superior olivary nucleus (SON), inferior colliculus (IC), and ipsilateral fiber tracts between the SON and IC delineated by tractography. RESULTS: We identified differences in fractional anisotropy of the SON between the SNHL cohort and controls (0.377 ± 0.056 vs. 0.422 ± 0.052; p = 0.009), but not in the IC. There were no differences in the mean diffusivity (MD) values in the IC and SON. Among younger children (≤5 yrs), MD was decreased in the SNHL cohort compared with controls in the IC (0.918 ± 0.051 vs. 1.120 ± 0.142; p < 0.001). However, among older children (>5 yrs), there were no differences in MD (1.124 ± 0.198 vs. 0.997 ± 0.103; p = 0.119). There were no differences in MD or fractional anisotropy in the white matter fibers of the IC-SON tract. CONCLUSIONS: Our results suggest abnormal neural tracts along the central auditory pathway among children with SNHL. Longitudinal studies should assess the prognostic value of these MRI-based findings for assessing long-term outcomes and determining intervention efficacy.
Asunto(s)
Sordera , Pérdida Auditiva Sensorineural , Sustancia Blanca , Humanos , Masculino , Niño , Adolescente , Lactante , Preescolar , Vías Auditivas/diagnóstico por imagen , Vías Auditivas/patología , Imagen de Difusión Tensora/métodos , Estudios Retrospectivos , Pérdida Auditiva Sensorineural/diagnóstico por imagen , Pérdida Auditiva Sensorineural/patología , Sordera/patología , Sustancia Blanca/diagnóstico por imagen , Tronco EncefálicoRESUMEN
OBJECTIVE: To determine the positivity rate of congenital cytomegalovirus (cCMV) testing among universal, hearing-targeted CMV testing (HT-cCMV) and delayed targeted dried blood spot (DBS) testing newborn screening programs, and to examine the characteristics of successful HT-cCMV testing programs. STUDY DESIGN: Prospective survey of birth hospitals performing early CMV testing. SETTING: Multiple institutions. METHODS: Birth hospitals participating in the National Institutes of Health ValEAR clinical trial were surveyed to determine the rates of cCMV positivity associated with 3 different testing approaches: universal testing, HT-cCMV, and DBS testing. A mixed methods model was created to determine associations between successful HT-cCMV screening and specific screening protocols. RESULTS: Eighty-two birth hospitals were surveyed from February 2019 to December 2021. Seven thousand six hundred seventy infants underwent universal screening, 9017 infants HT-cCMV and 535 infants delayed DBS testing. The rates of cCMV positivity were 0.5%, 1.5%, and 7.3%, respectively. The positivity rate for universal CMV screening was less during the COVID-19 pandemic than that reported prior to the pandemic. There were no statistically significant drops in positivity for any approach during the pandemic. For HT-cCMV testing, unique order sets and rigorous posttesting protocols were associated with successful screening programs. CONCLUSION: Rates of cCMV positivity differed among the 3 approaches. The rates are comparable to cohort studies reported in the literature. Universal CMV prevalence decreased during the pandemic but not significantly. Institutions with specific order set for CMV testing where the primary care physician orders the test and the nurse facilitates the testing process exhibited higher rates of HT-cCMV testing.
Asunto(s)
Infecciones por Citomegalovirus , Tamizaje Neonatal , Humanos , Infecciones por Citomegalovirus/diagnóstico , Infecciones por Citomegalovirus/congénito , Infecciones por Citomegalovirus/epidemiología , Tamizaje Neonatal/métodos , Recién Nacido , Estudios Prospectivos , COVID-19/epidemiología , COVID-19/diagnóstico , Estados Unidos/epidemiología , Pruebas con Sangre Seca , Femenino , MasculinoRESUMEN
Cochlear hair cells convert sound into electrical signals that are relayed via the spiral ganglion neurons to the central auditory pathway. Hair cells are vulnerable to damage caused by excessive noise, aging, and ototoxic agents. Non-mammals can regenerate lost hair cells by mitotic regeneration and direct transdifferentiation of surrounding supporting cells. However, in mature mammals, damaged hair cells are not replaced, resulting in permanent hearing loss. Recent studies have uncovered mechanisms by which sensory organs in non-mammals and the neonatal mammalian cochlea regenerate hair cells, and outlined possible mechanisms why this ability declines rapidly with age in mammals. Here, we review similarities and differences between avian, zebrafish, and mammalian hair cell regeneration. Moreover, we discuss advances and limitations of hair cell regeneration in the mature cochlea and their potential applications to human hearing loss.
RESUMEN
OBJECTIVES: Assess the relationship between public interest in ankyloglossia as determined by internet search volume and real-world medical claims data. STUDY DESIGN: Retrospective Cohort Study. SETTING: This retrospective cohort study was conducted using claims data from the Merative™ Marketscan® Research Databases. The internet search data was collected from Google Trends. METHODS: Annual Google Trends data were compiled using search terms associated with "ankyloglossia" and "frenotomy" for the years 2011 to 2021. We obtained incidence of ankyloglossia diagnoses and frenotomy procedures in children under 12 months from Marketscan relative to all infants enrolled. We compared associations between search and incidence data among US states and over time. RESULTS: Google search correlated with ankyloglossia incidence (r = 0.4104, P = .0031) and with frenotomy incidence (r = 0.4062, P = .0034) per state. Ankyloglossia diagnoses increased with Google search index (coefficient = 0.336, 95% confidence interval [CI] 0.284, 0.388) and year (coefficient = 0.028, 95% CI 0.025, 0.031). Similarly, frenotomy procedures increased with Google search index (coefficient = 0.371, 95% CI 0.313, 0.429) and year (coefficient = 0.027, 95% CI 0.024, 0.030). CONCLUSIONS: Associations between online ankyloglossia search trends and both diagnosis and treatment rates, persist across US regions and timeframes. Internet search trends are pivotal in shaping pediatric health care decisions, driving clinical consensus, and disseminating evidence-based information.
Asunto(s)
Anquiloglosia , Humanos , Anquiloglosia/epidemiología , Anquiloglosia/cirugía , Estudios Retrospectivos , Lactante , Estados Unidos , Femenino , Internet , Masculino , Incidencia , Recién Nacido , Bases de Datos FactualesRESUMEN
Cochlear melanocytes are intermediate cells in the stria vascularis that generate endocochlear potentials required for auditory function. Human PAX3 mutations cause Waardenburg syndrome and abnormalities of skin and retinal melanocytes, manifested as congenital hearing loss (~ 70%) and hypopigmentation of skin, hair and eyes. However, the underlying mechanism of hearing loss remains unclear. Cochlear melanocytes in the stria vascularis originated from Pax3-traced melanoblasts and Plp1-traced Schwann cell precursors, both of which derive from neural crest cells. Here, using a Pax3-Cre knock-in mouse that allows lineage tracing of Pax3-expressing cells and disruption of Pax3, we found that Pax3 deficiency causes foreshortened cochlea, malformed vestibular apparatus, and neural tube defects. Lineage tracing and in situ hybridization show that Pax3+ derivatives contribute to S100+, Kir4.1+ and Dct+ melanocytes (intermediate cells) in the developing stria vascularis, all of which are significantly diminished in Pax3 mutant animals. Taken together, these results suggest that Pax3 is required for the development of neural crest cell-derived cochlear melanocytes, whose absence may contribute to congenital hearing loss of Waardenburg syndrome in humans.
Asunto(s)
Sordera , Pérdida Auditiva Sensorineural , Pérdida Auditiva , Síndrome de Waardenburg , Ratones , Animales , Humanos , Síndrome de Waardenburg/genética , Cóclea , Estría Vascular , Pérdida Auditiva Sensorineural/genética , Melanocitos , Factor de Transcripción PAX3/genéticaRESUMEN
BACKGROUND: The Micra AV Coverage with Evidence Development study is a novel analysis of utilization and outcomes associated with Micra AV leadless pacing in US Medicare patients. OBJECTIVE: The purpose of this study was to describe patient characteristics, complications, and outcomes of patients implanted with a Micra AV leadless pacemaker compared with a contemporaneous cohort of patients implanted with a dual chamber transvenous pacemaker. METHODS: Patients implanted with Micra AV (n = 7471) or a dual chamber transvenous pacemaker (n = 107,800) from February 5, 2020, through December 1, 2021, were identified using device registry-linked Medicare claims data. Acute complications were assessed at 30 days, and chronic complications, reinterventions, and all-cause mortality were assessed at 6 months. RESULTS: Patients implanted with Micra AV had higher rates of end-stage renal disease (14.9% vs 2.0%; P < .0001) and overall comorbidity burden (mean Charlson Comorbidity Index 4.9 vs 3.8; P < .0001). There was no difference in the unadjusted rate of complications at 30 days (9.1% vs 8.7%; P = .61), and patients implanted with Micra AV had a significantly lower adjusted rate of complications (8.6% vs 11.0%; P < .0001). At 6 months, patients implanted with Micra AV had significantly lower rates of complications (adjusted hazard ratio 0.50; 95% confidence interval 0.43-0.57; P < .0001) and reinterventions (adjusted hazard ratio 0.46; 95% confidence interval 0.36-0.58; P < .0001). Patients implanted with Micra AV had higher all-cause mortality at 30 days and 6 months, likely because of differences in the underlying risk of mortality. CONCLUSION: Patients implanted with Micra AV had similar rates of complications at 30 days and significantly lower rates of complications and reinterventions at 6 months, despite being sicker than patients implanted with a transvenous pacemaker.
Asunto(s)
Medicare , Marcapaso Artificial , Estados Unidos/epidemiología , Humanos , Anciano , Resultado del Tratamiento , Diseño de Equipo , Marcapaso Artificial/efectos adversos , Prótesis e Implantes , Estimulación Cardíaca Artificial/efectos adversosRESUMEN
Adeno-associated virus (AAV)-mediated gene transfer has shown promise in rescuing mouse models of genetic hearing loss, but how viral capsid and promoter selection affects efficacy is poorly characterized. Here, we tested combinations of AAVs and promoters to deliver Tmprss3, mutations in which are associated with hearing loss in humans. Tmprss3tm1/tm1 mice display severe cochlear hair cell degeneration, loss of auditory brainstem responses, and delayed loss of spiral ganglion neurons. Under the ubiquitous CAG promoter and AAV-KP1 capsid, Tmprss3 overexpression caused striking cytotoxicity in vitro and in vivo and failed to rescue degeneration or dysfunction of the Tmprss3tm1/tm1 cochlea. Reducing the dosage or using AAV-DJ-CAG-Tmprss3 diminished cytotoxicity without rescue of the Tmprss3tm1/tm1 cochlea. Finally, the combination of AAV-KP1 capsid and the EF1α promoter prevented cytotoxicity and reduced hair cell degeneration, loss of spiral ganglion neurons, and improved hearing thresholds in Tmprss3tm1/tm1 mice. Together, our study illustrates toxicity of exogenous genes and factors governing rescue efficiency, and suggests that cochlear gene therapy likely requires precisely targeted transgene expression.
RESUMEN
Coronaviruses have been the causative agent of three epidemics and pandemics in the past two decades, including the ongoing COVID-19 pandemic. A broadly-neutralizing coronavirus therapeutic is desirable not only to prevent and treat COVID-19, but also to provide protection for high-risk populations against future emergent coronaviruses. As all coronaviruses use spike proteins on the viral surface to enter the host cells, and these spike proteins share sequence and structural homology, we set out to discover cross-reactive biologic agents targeting the spike protein to block viral entry. Through llama immunization campaigns, we have identified single domain antibodies (VHHs) that are cross-reactive against multiple emergent coronaviruses (SARS-CoV, SARS-CoV-2, and MERS). Importantly, a number of these antibodies show sub-nanomolar potency towards all SARS-like viruses including emergent CoV-2 variants. We identified nine distinct epitopes on the spike protein targeted by these VHHs. Further, by engineering VHHs targeting distinct, conserved epitopes into multi-valent formats, we significantly enhanced their neutralization potencies compared to the corresponding VHH cocktails. We believe this approach is ideally suited to address both emerging SARS-CoV-2 variants during the current pandemic as well as potential future pandemics caused by SARS-like coronaviruses.
Asunto(s)
COVID-19 , Camélidos del Nuevo Mundo , Anticuerpos de Dominio Único , Humanos , Animales , SARS-CoV-2 , Glicoproteína de la Espiga del Coronavirus , Pandemias , EpítoposRESUMEN
Identification of favorable biophysical properties for protein therapeutics as part of developability assessment is a crucial part of the preclinical development process. Successful prediction of such properties and bioassay results from calculated in silico features has potential to reduce the time and cost of delivering clinical-grade material to patients, but nevertheless has remained an ongoing challenge to the field. Here, we demonstrate an automated and flexible machine learning workflow designed to compare and identify the most powerful features from computationally derived physiochemical feature sets, generated from popular commercial software packages. We implement this workflow with medium-sized datasets of human and humanized IgG molecules to generate predictive regression models for two key developability endpoints, hydrophobicity and poly-specificity. The most important features discovered through the automated workflow corroborate several previous literature reports, and newly discovered features suggest directions for further research and potential model improvement.
Asunto(s)
Anticuerpos Monoclonales , Inmunoglobulina G , Humanos , Anticuerpos Monoclonales/química , Aprendizaje AutomáticoRESUMEN
The mammalian cochlea is composed of sensory hair cells as well as multiple different types of non-sensory supporting cells. Pillar cells are one type of supporting cell that form the tunnel of Corti and include two morphologically and functionally distinct subtypes: inner pillar cells (IPCs) and outer pillar cells (OPCs). The processes of specification and differentiation of inner versus outer pillar cells are still unclear. Here, we show that ß-Catenin is required for establishing IPC identity in the mammalian cochlea. To differentiate the transcriptional and adhesion roles of ß-Catenin in establishing IPC identity, we examined two different models of ß-Catenin deletion; one that deletes both transcriptional and structural functions and one which retains cell adhesion function but lacks transcriptional function. Here, we show that cochleae lacking ß-Catenin transcriptional function lost IPCs and displayed extranumerary OPCs, indicating its requirement for establishing IPC identity. Overexpression of ß-Catenin induced proliferation within IPCs but not ectopic IPCs. Single-cell transcriptomes of supporting cells lacking ß-Catenin transcriptional function show a loss of the IPC and gain of OPC signatures. Finally, targeted deletion of ß-Catenin in IPCs also led to the loss of IPC identity, indicating a cell autonomous role of ß-Catenin in establishing IPC identity. As IPCs have the capacity to regenerate sensory hair cells in the postnatal cochlea, our results will aid in future IPC-based hair cell regeneration strategies.