The long tail and rare disease research: the impact of next-generation sequencing for rare Mendelian disorders

A generalist medical language model for disease diagnosis assistance

The delivery of accurate diagnoses is crucial in healthcare and represents the gateway to appropriate and timely treatment. Although recent large language models (LLMs) have demonstrated impressive capabilities in few-shot or zero-shot learning, their effectiveness in clinical diagnosis remains unproven. Here we present MedFound, a generalist medical language model with 176 billion parameters, pre-trained on a large-scale corpus derived from diverse medical text and real-world clinical records. We further fine-tuned MedFound to learn physicians' inferential diagnosis with a self-bootstrapping strategy-based chain-of-thought approach and introduced a unified preference alignment framework to align it with standard clinical practice. Extensive experiments demonstrate that our medical LLM outperforms other baseline LLMs and specialized models in in-distribution (common diseases), out-of-distribution (external validation) and long-tailed distribution (rare diseases) scenarios across eight specialties. Further ablation studies indicate the effectiveness of key components in our medical LLM training approach. We conducted a comprehensive evaluation of the clinical applicability of LLMs for diagnosis involving artificial intelligence (AI) versus physician comparison, AI-assistance study and human evaluation framework. Our proposed framework incorporates eight clinical evaluation metrics, covering capabilities such as medical record summarization, diagnostic reasoning and risk management. Our findings demonstrate the model's feasibility in assisting physicians with disease diagnosis as part of the clinical workflow.

Epigenetic disorders: Lessons from the animals–animal models in chromatinopathies

Chromatinopathies are defined as genetic disorders caused by mutations in genes coding for protein involved in the chromatin state balance. So far 82 human conditions have been described belonging to this group of congenital disorders, sharing some molecular features and clinical signs. For almost all of these conditions, no specific treatment is available. For better understanding the molecular cascade caused by chromatin imbalance and for envisaging possible therapeutic strategies it is fundamental to combine clinical and basic research studies. To this end, animal modelling systems represent an invaluable tool to study chromatinopathies. In this review, we focused on available data in the literature of animal models mimicking the human genetic conditions. Importantly, affected organs and abnormalities are shared in the different animal models and most of these abnormalities are reported as clinical manifestation, underlying the parallelism between clinics and translational research.

Agreement and Reliability Analysis of Machine Learning Scaling and Wireless Monitoring in the Assessment of Acute Proximal Weakness by Experts and Non-Experts: A Feasibility Study

Assessing the symptoms of proximal weakness caused by neurological deficits requires the knowledge and experience of neurologists. Recent advances in machine learning and the Internet of Things have resulted in the development of automated systems that emulate physicians’ assessments. The application of those systems requires not only accuracy in the classification but also reliability regardless of users’ proficiency in the real environment for the clinical point-of-care and the personalized health management. This study provides an agreement and reliability analysis of using a machine learning-based scaling of Medical Research Council (MRC) proximal scores to evaluate proximal weakness by experts and non-experts. The system trains an ensemble learning model using the signals from sensors attached to the limbs of patients in a neurological intensive care unit. For the agreement analysis, we investigated the percent agreement of MRC proximal scores and Bland-Altman plots of kinematic features between the expert- and non-expert scaling. We also analyzed the intra-class correlation coefficients (ICCs) of kinematic features and Krippendorff’s alpha of the observers’ scaling for the reliability analysis. The mean percent agreement between the expert- and the non-expert scaling was 0.542 for manual scaling and 0.708 for autonomous scaling. The ICCs of kinematic features measured using sensors ranged from 0.742 to 0.850, whereas the Krippendorff’s alpha of manual scaling for the three observers was 0.275. The autonomous assessment system can be utilized by the caregivers, paramedics, or other observers during an emergency to evaluate acute stroke patients.

Finding commonalities in rare diseases through the undiagnosed diseases network

Objective

When studying any specific rare disease, heterogeneity and scarcity of affected individuals has historically hindered investigators from discerning on what to focus to understand and diagnose a disease. New nongenomic methodologies must be developed that identify similarities in seemingly dissimilar conditions.

Materials and Methods

This observational study analyzes 1042 patients from the Undiagnosed Diseases Network (2015-2019), a multicenter, nationwide research study using phenotypic data annotated by specialized staff using Human Phenotype Ontology terms. We used Louvain community detection to cluster patients linked by Jaccard pairwise similarity and 2 support vector classifier to assign new cases. We further validated the clusters’ most representative comorbidities using a national claims database (67 million patients).

Results

Patients were divided into 2 groups: those with symptom onset before 18 years of age (n = 810) and at 18 years of age or older (n = 232) (average symptom onset age: 10 [interquartile range, 0-14] years). For 810 pediatric patients, we identified 4 statistically significant clusters. Two clusters were characterized by growth disorders, and developmental delay enriched for hypotonia presented a higher likelihood of diagnosis. Support vector classifier showed 0.89 balanced accuracy (0.83 for Human Phenotype Ontology terms only) on test data.

Discussions

To set the framework for future discovery, we chose as our endpoint the successful grouping of patients by phenotypic similarity and provide a classification tool to assign new patients to those clusters.

Conclusion

This study shows that despite the scarcity and heterogeneity of patients, we can still find commonalities that can potentially be harnessed to uncover new insights and targets for therapy.

The future of genomics in Ireland - focus on genomics for health

Genomics is revolutionizing biomedical research, medicine and healthcare globally in academic, public and industry sectors alike. Concrete examples around the world show that huge benefits for patients, society and economy can be accrued through effective and responsible genomic research and clinical applications. Unfortunately, Ireland has fallen behind and needs to act now in order to catch up. Here, we identify key issues that have resulted in Ireland lagging behind, describe how genomics can benefit Ireland and its people and outline the measures needed to make genomics work for Ireland and Irish patients. There is now an urgent need for a national genomics strategy that enables an effective, collaborative, responsible, well-regulated, and patient centred environment where genome research and clinical genomics can thrive.  We present eight recommendations that could be the pillars of a national genomics health strategy.

Genetic determinants of fracture non-union: A systematic review from the literature

Approximately 10-15% of fracture patients suffer impaired healing, which is either delayed or even results in non-union. We performed a Systematic Review, aiming to examine the types and frequency of specific genetic abnormalities in patients experiencing bone fracture and to ascertain whether a genetic association exists regarding the tendency for some patients to suffer fracture non-union or postoperative non-union events. GO and KEGG analyses were used to identify the likely function of the genes involved. Furthermore, we evaluated the functional significance of single nucleotide polymorphisms using RegulomeDB and GTEx. Seven eligible studies involving 29 genes and 89 SNPs were analyzed in this review. We found that the polymorphisms in gene NOS2, NOG, BMP4, CYR61, IL1β and FGFR1 apparently predisposed patients to fracture non-union, while the polymorphisms in gene MMP13, BMP6 and FAM5C appeared to provide protection from non-union. Bioinformatics analysis suggested that these genes were enriched in inflammatory pathways, suggesting that inflammation may be a potential factor involved in fracture non-union. Three SNPs (rs17563, rs3753793 and rs2853550) had smaller RegulomeDB scores, indicating significant biological function. In conclusion, we have identified a number of genes and their polymorphisms that might contribute to a genetic susceptibility to fracture non-union. Further studies with larger cohorts will enhance our understanding of fracture non-union and may inform and direct early interventions.

RegulationSpotter: annotation and interpretation of extratranscriptic DNA variants

RegulationSpotter is a web-based tool for the user-friendly annotation and interpretation of DNA variants located outside of protein-coding transcripts (extratranscriptic variants). It is designed for clinicians and researchers who wish to assess the potential impact of the considerable number of non-coding variants found in Whole Genome Sequencing runs. It annotates individual variants with underlying regulatory features in an intuitive way by assessing over 100 genome-wide annotations. Additionally, it calculates a score, which reflects the regulatory potential of the variant region. Its dichotomous classifications, ‘functional’ or ‘non-functional’, and a human-readable presentation of the underlying evidence allow a biologically meaningful interpretation of the score. The output shows key aspects of every variant and allows rapid access to more detailed information about its possible role in gene regulation. RegulationSpotter can either analyse single variants or complete VCF files. Variants located within protein-coding transcripts are automatically assessed by MutationTaster as well as by RegulationSpotter to account for possible intragenic regulatory effects. RegulationSpotter offers the possibility of using phenotypic data to focus on known disease genes or genomic elements interacting with them. RegulationSpotter is freely available at https://www.regulationspotter.org.

The combination of whole-exome sequencing and clinical analysis allows better diagnosis of rare syndromic retinal dystrophies

PURPOSE

To identify the accurate clinical diagnosis of rare syndromic inherited retinal diseases (IRDs) based on the combination of clinical and genetic analyses.

METHODS

Four unrelated families with various autosomal recessive syndromic inherited retinal diseases were genetically investigated using whole-exome sequencing (WES).

RESULTS

Two affected subjects in family MOL0760 presented with a distinctive combination of short stature, developmental delay, congenital mental retardation, microcephaly, facial dysmorphism and retinitis pigmentosa (RP). Subjects were clinically diagnosed with suspected Kabuki syndrome. WES revealed a homozygous nonsense mutation (c.5492dup, p.Asn1831Lysfs*8) in VPS13B that is known to cause Cohen syndrome. The index case of family MOL1514 presented with both RP and liver dysfunction, suspected initially to be related. WES identified a homozygous frameshift mutation (c.1787_1788del, p.His596Argfs*47) in AGBL5, associated with nonsyndromic RP. The MOL1592 family included three affected subjects with crystalline retinopathy, skin ichthyosis, short stature and congenital adrenal hypoplasia, and were found to harbour a homozygous nonsense mutation (c.682C>T, p.Arg228Cys) in ALDH3A2, reported to cause Sjögren-Larsson syndrome (SLS). In the fourth family, SJ002, two siblings presented with hypotony, psychomotor delay, dysmorphic facial features, pathologic myopia, progressive external ophthalmoplegia and diffuse retinal atrophy. Probands were suspected to have atypical Kearns-Sayre syndrome, but were diagnosed with combined oxidative phosphorylation deficiency-20 due to a novel suspected missense variant (c.1691C>T, p.Ala564Val) in VARS2.

CONCLUSION

Our findings emphasize the important complement of WES and thorough clinical investigation in establishing precise clinical diagnosis. This approach constitutes the basis for personalized medicine in rare IRDs.

GenePy - a score for estimating gene pathogenicity in individuals using next-generation sequencing data

BACKGROUND

Next-generation sequencing is revolutionising diagnosis and treatment of rare diseases, however its application to understanding common disease aetiology is limited. Rare disease applications binarily attribute genetic change(s) at a single locus to a specific phenotype. In common diseases, where multiple genetic variants within and across genes contribute to disease, binary modelling cannot capture the burden of pathogenicity harboured by an individual across a given gene/pathway. We present GenePy, a novel gene-level scoring system for integration and analysis of next-generation sequencing data on a per-individual basis that transforms NGS data interpretation from variant-level to gene-level. This simple and flexible scoring system is intuitive and amenable to integration for machine learning, network and topological approaches, facilitating the investigation of complex phenotypes.

RESULTS

Whole-exome sequencing data from 508 individuals were used to generate GenePy scores. For each variant a score is calculated incorporating: i) population allele frequency estimates; ii) individual zygosity, determined through standard variant calling pipelines and; iii) any user defined deleteriousness metric to inform on functional impact. GenePy then combines scores generated for all variants observed into a single gene score for each individual. We generated a matrix of ~ 14,000 GenePy scores for all individuals for each of sixteen popular deleteriousness metrics. All per-gene scores are corrected for gene length. The majority of genes generate GenePy scores < 0.01 although individuals harbouring multiple rare highly deleterious mutations can accumulate extremely high GenePy scores. In the absence of a comparator metric, we examine GenePy performance in discriminating genes known to be associated with three common, complex diseases. A Mann-Whitney U test conducted on GenePy scores for this positive control gene in cases versus controls demonstrates markedly more significant results (p = 1.37 × 10- 4) compared to the most commonly applied association tool that combines common and rare variation (p = 0.003).

CONCLUSIONS

Per-gene per-individual GenePy scores are intuitive when assessing genetic variation in individual patients or comparing scores between groups. GenePy outperforms the currently accepted best practice tools for combining common and rare variation. GenePy scores are suitable for downstream data integration with transcriptomic and proteomic data that also report at the gene level.

Clinical utility in infants with suspected monogenic conditions through next-generation sequencing

Background

Rare diseases are complex disorders with huge variability in clinical manifestations. Decreasing cost of next‐generation sequencing (NGS) tests in recent years made it affordable. We witnessed the diagnostic yield and clinical use of different NGS strategies on a myriad of monogenic disorders in a pediatric setting.

Methods

Next‐generation sequencing tests are performed for 98 unrelated Chinese patients within their first year of life, who were admitted to Xin Hua Hospital, affiliated with Shanghai Jiao Tong University School of Medicine, during a 2‐year period.

Results

Clinical indications for NGS tests included a range of medical concerns. The mean age was 4.4 ± 4.2 months of age for infants undergoing targeting specific (known) disease‐causing genes (TRS) analysis, and 4.4 ± 4.3 months of age for whole‐exome sequencing (WES) (p > 0.05). A molecular diagnosis is done in 72 infants (73.47%), which finds a relatively high yield with phenotypes of metabolism/homeostasis abnormality (HP: 0001939) (odds ratio, 1.83; 95% CI, 0.56–6.04; p = 0.32) and a significantly low yield with atypical symptoms (without a definite HPO term) (odds ratio, 0.08; 95% CI, 0.01–0.73; p = 0.03). TRS analysis provides molecular yields higher than WES (p = 0.01). Ninety‐eight different mutations are discovered in 72 patients. Twenty‐seven of them have not been reported previously. Nearly half (43.06%, 31/72) of the patients are found to carry 11 common disorders, mostly being inborn errors of metabolism (IEM) and neurogenetic disorders and all of them are observed through TRS analysis. Eight positive cases are identified through WES, and all of them are sporadic, of highly variable phenotypes and severity. There are 26 patients with negative findings in this study.

Conclusion

This study provides evidence that NGS can yield high success rates in a tertiary pediatric setting, but suggests that the scope of known Mendelian conditions may be considerably broader than currently recognized.

Use of zebrafish models to investigate rare human disease

Rare diseases are collectively common and often extremely debilitating. Following the emergence of next-generation sequencing (NGS) technologies, the variants underpinning rare genetic disorders are being unearthed at an accelerating rate. However, many rare conditions lack effective treatments due to their poorly understood pathophysiology. There is therefore a growing demand for the development of novel experimental models of rare genetic diseases, so that potentially causative variants can be validated, pathogenic mechanisms can be investigated and therapeutic targets can be identified. Animal models of rare diseases need to be genetically and physiologically similar to humans, and well-suited to large-scale experimental manipulation, considering the vast number of novel variants that are being identified through NGS. The zebrafish has emerged as a popular model system for investigating these variants, combining conserved vertebrate characteristics with a capacity for large-scale phenotypic and therapeutic screening. In this review, we aim to highlight the unique advantages of the zebrafish over other in vivo model systems for the large-scale study of rare genetic variants. We will also consider the generation of zebrafish disease models from a practical standpoint, by discussing how genome editing technologies, particularly the recently developed clustered regularly interspaced repeat (CRISPR)/CRISPR-associated protein 9 system, can be used to model rare pathogenic variants in zebrafish. Finally, we will review examples in the literature where zebrafish models have played a pivotal role in confirming variant causality and revealing the underlying mechanisms of rare diseases, often with wider implications for our understanding of human biology.

Molecular diagnosis of coenzyme Q10 deficiency: an update

INTRODUCTION

Coenzyme Q₁₀ (CoQ) deficiency syndromes comprise a growing number of genetic disorders. While primary CoQ deficiency syndromes are rare diseases, secondary deficiencies have been related to both genetic and environmental conditions, which are the main causes of biochemical CoQ deficiency. The diagnosis is the essential first step for planning future treatment strategies, as the potential treatability of CoQ deficiency is the most critical issue for the patients. Areas covered: While the quickest and most effective tool to define a CoQ-deficient status is its biochemical determination in biological fluids or tissues, this quantification does not provide a definite diagnosis of a CoQ-deficient status nor insight about the genetic etiology of the disease. The different laboratory tests to check for CoQ deficiency are evaluated in order to choose the best diagnostic pathway for the patient. Expert commentary: New insights are being discovered about the implication of new proteins in the intricate CoQ biosynthetic pathway. These insights reinforce the idea that next generation sequencing diagnostic strategies are the unique alternative in terms of rapid and accurate molecular diagnosis of CoQ deficiency.

Neurodevelopmental Genetic Diseases Associated With Microdeletions and Microduplications of Chromosome 17p13.3

Chromosome 17p13.3 is a region of genomic instability that is linked to different rare neurodevelopmental genetic diseases, depending on whether a deletion or duplication of the region has occurred. Chromosome microdeletions within 17p13.3 can result in either isolated lissencephaly sequence (ILS) or Miller-Dieker syndrome (MDS). Both conditions are associated with a smooth cerebral cortex, or lissencephaly, which leads to developmental delay, intellectual disability, and seizures. However, patients with MDS have larger deletions than patients with ILS, resulting in additional symptoms such as poor muscle tone, congenital anomalies, abnormal spasticity, and craniofacial dysmorphisms. In contrast to microdeletions in 17p13.3, recent studies have attracted considerable attention to a condition known as a 17p13.3 microduplication syndrome. Depending on the genes involved in their microduplication, patients with 17p13.3 microduplication syndrome may be categorized into either class I or class II. Individuals in class I have microduplications of the YWHAE gene encoding 14-3-3ε, as well as other genes in the region. However, the PAFAH1B1 gene encoding LIS1 is never duplicated in these patients. Class I microduplications generally result in learning disabilities, autism, and developmental delays, among other disorders. Individuals in class II always have microduplications of the PAFAH1B1 gene, which may include YWHAE and other genetic microduplications. Class II microduplications generally result in smaller body size, developmental delays, microcephaly, and other brain malformations. Here, we review the phenotypes associated with copy number variations (CNVs) of chromosome 17p13.3 and detail their developmental connection to particular microdeletions or microduplications. We also focus on existing single and double knockout mouse models that have been used to study human phenotypes, since the highly limited number of patients makes a study of these conditions difficult in humans. These models are also crucial for the study of brain development at a mechanistic level since this cannot be accomplished in humans. Finally, we emphasize the usefulness of the CRISPR/Cas9 system and next generation sequencing in the study of neurodevelopmental diseases.

Drugs for rare disorders

Estimates of the frequencies of rare disorders vary from country to country; the global average defined prevalence is 40 per 100 000 (0.04%). Some occur in only one or a few patients. However, collectively rare disorders are fairly common, affecting 6-8% of the US population, or about 30 million people, and a similar number in the European Union. Most of them affect children and most are genetically determined. Diagnosis can be difficult, partly because of variable presentations and partly because few clinicians have experience of individual rare disorders, although they may be assisted by searching databases. Relatively few rare disorders have specific pharmacological treatments (so-called orphan drugs), partly because of difficulties in designing trials large enough to determine benefits and harms alike. Incentives have been introduced to encourage the development of orphan drugs, including tax credits and research aids, simplification of marketing authorization procedures and exemption from fees, and extended market exclusivity. Consequently, the number of applications for orphan drugs has grown, as have the costs of using them, so much so that treatments may not be cost-effective. It has therefore been suggested that not-for-profit organizations that are socially motivated to reduce those costs should be tasked with producing them. A growing role for patient organizations, improved clinical and translational infrastructures, and developments in genetics have also contributed to successful drug development. The translational discipline of clinical pharmacology is an essential component in drug development, including orphan drugs. Clinical pharmacologists, skilled in basic pharmacology and its links to clinical medicine, can be involved at all stages. They can contribute to the delineation of genetic factors that determine clinical outcomes of pharmacological interventions, develop biomarkers, design and perform clinical trials, assist regulatory decision making, and conduct postmarketing surveillance and pharmacoepidemiological and pharmacoeconomic assessments.

Whole-Genome Sequencing of a Canine Family Trio Reveals a FAM83G Variant Associated with Hereditary Footpad Hyperkeratosis

Over 250 Mendelian traits and disorders, caused by rare alleles have been mapped in the canine genome. Although each disease is rare in the dog as a species, they are collectively common and have major impact on canine health. With SNP-based genotyping arrays, genome-wide association studies (GWAS) have proven to be a powerful method to map the genomic region of interest when 10–20 cases and 10–20 controls are available. However, to identify the genetic variant in associated regions, fine-mapping and targeted resequencing is required. Here we present a new approach using whole-genome sequencing (WGS) of a family trio without prior GWAS. As a proof-of-concept, we chose an autosomal recessive disease known as hereditary footpad hyperkeratosis (HFH) in Kromfohrländer dogs. To our knowledge, this is the first time this family trio WGS-approach has been used successfully to identify a genetic variant that perfectly segregates with a canine disorder. The sequencing of three Kromfohrländer dogs from a family trio (an affected offspring and both its healthy parents) resulted in an average genome coverage of 9.2X per individual. After applying stringent filtering criteria for candidate causative coding variants, 527 single nucleotide variants (SNVs) and 15 indels were found to be homozygous in the affected offspring and heterozygous in the parents. Using the computer software packages ANNOVAR and SIFT to functionally annotate coding sequence differences, and to predict their functional effect, resulted in seven candidate variants located in six different genes. Of these, only FAM83G:c155G > C (p.R52P) was found to be concordant in eight additional cases, and 16 healthy Kromfohrländer dogs.