Deep neural networks for phenotype prediction in rare diseases

deAndrés-Galiana, Enrique J.; Fernández-Ovies, Francisco Javier; Cernea, Ana; Fernández-Martínez, Juan Luis; Kloczkowski, Andrzej

doi:10.1016/b978-0-12-817133-2.00008-2

Cited by 4 publications

(2 citation statements)

References 27 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…At the inflammatory level, IBM fibroblast transcriptome revealed altered inflammatory genes resembling muscle features, including MHC1 ( HLA genes) and IFN genes. These alterations were found in an IBM deep neural network model 29,30 . At the functional scale, cytokines secretion in IBM fibroblasts was increased.…”

Section: Discussionmentioning

confidence: 84%

“…These alterations were found in an IBM deep neural network model. 29 , 30 At the functional scale, cytokines secretion in IBM fibroblasts was increased. Previous studies in IBM muscle detected CD8+ T cell‐secreted cytokines and chemokines related to IFN‐ɣ genes ( CXCL9 , CXCL10 , CCL5 , and CCL18 ) and others, including TNF , IL7 , IL15 , IL16 , and IL32 , 1 , 26 some of which were also detected in the sera of IBM patients.…”

Section: Discussionmentioning

confidence: 99%

See 1 more Smart Citation

Unravelling inclusion body myositis using a patient‐derived fibroblast model

Cantó-Santos

Valls-Roca

Tobías

et al. 2023

J cachexia sarcopenia muscle

View full text Add to dashboard Cite

Background Inclusion body myositis (IBM) is an inflammatory myopathy clinically characterized by proximal and distal muscle weakness, with inflammatory infiltrates, rimmed vacuoles and mitochondrial changes in muscle histopathology. There is scarce knowledge on IBM aetiology, and non‐established biomarkers or effective treatments are available, partly due to the lack of validated disease models. Methods We have performed transcriptomics and functional validation of IBM muscle pathological hallmarks in fibroblasts from IBM patients (n = 14) and healthy controls (n = 12), paired by age and sex. The results comprise an mRNA‐seq, together with functional inflammatory, autophagy, mitochondrial and metabolic changes between patients and controls. Results Gene expression profile of IBM vs control fibroblasts revealed 778 differentially expressed genes (P‐value adj < 0.05) related to inflammation, mitochondria, cell cycle regulation and metabolism. Functionally, an increased inflammatory profile was observed in IBM fibroblasts with higher supernatant cytokine secretion (three‐fold increase). Autophagy was reduced considering basal protein mediators (18.4% reduced), time‐course autophagosome formation (LC3BII 39% reduced, P‐value < 0.05), and autophagosome microscopic evaluation. Mitochondria displayed reduced genetic content (by 33.9%, P‐value < 0.05) and function (30.2%‐decrease in respiration, 45.6%‐decline in enzymatic activity (P‐value < 0.001), 14.3%‐higher oxidative stress, 135.2%‐increased antioxidant defence (P‐value < 0.05), 11.6%‐reduced mitochondrial membrane potential (P‐value < 0.05) and 42.8%‐reduced mitochondrial elongation (P‐value < 0.05)). In accordance, at the metabolite level, organic acid showed a 1.8‐fold change increase, with conserved amino acid profile. Correlating to disease evolution, oxidative stress and inflammation emerge as potential markers of prognosis. Conclusions These findings confirm the presence of molecular disturbances in peripheral tissues from IBM patients and prompt patients' derived fibroblasts as a promising disease model, which may eventually be exported to other neuromuscular disorders. We additionally identify new molecular players in IBM associated with disease progression, setting the path to deepen in disease aetiology, in the identification of novel biomarkers or in the standardization of biomimetic platforms to assay new therapeutic strategies for preclinical studies.

show abstract

Section: Discussionmentioning

confidence: 84%

Section: Discussionmentioning

confidence: 99%

Unravelling inclusion body myositis using a patient‐derived fibroblast model

Cantó-Santos

Valls-Roca

Tobías

et al. 2023

J cachexia sarcopenia muscle

View full text Add to dashboard Cite

show abstract

Prediction of Functional Effects of Protein Amino Acid Mutations

Álvarez-Machancoses

Faraggi

Andrés-Galiana

et al. 2023

Bioinformatics and Biomedical Engineering

View full text Add to dashboard Cite

Prediction of Deleterious Single Amino Acid Polymorphisms with a Consensus Holdout Sampler

Álvarez-Machancoses,

Faraggi,

deAndrés-Galiana

et al. 2024

View full text Add to dashboard Cite

Background: Single Amino Acid Polymorphisms (SAPs) or nonsynonymous Single Nucleotide Variants (nsSNVs) are the most common genetic variations. They result from missense mutations where a single base pair substitution changes the genetic code in such a way that the triplet of bases (codon) at a given position is coding a different amino acid. Since genetic mutations sometimes cause genetic diseases, it is important to comprehend and foresee which variations are harmful and which ones are neutral (not causing changes in the phenotype). This can be posed as a classification problem. Methods: Computational methods using machine intelligence are gradually replacing repetitive and exceedingly overpriced mutagenic tests. By and large, uneven quality, deficiencies, and irregularities of nsSNVs datasets debase the convenience of artificial intelligence-based methods. Subsequently, strong and more exact approaches are needed to address these problems. In the present work paper, we show a consensus classifier built on the holdout sampler, which appears strong and precise and outflanks all other popular methods. objective: Roughly a half of known disease-related mutations are due to non-synonymous variants [8-9], expressed as amino-acid mutations. Therefore, it is important to unravel the links between nonsynonymous Single Nucleotide Variants and associated diseases to discriminate between pathogenic and neutral substitutions. It has been found that these substitutions could be directly related to pathological effects such as Parkinson’s or Alzheimer’s diseases, or to the involvement in complex diseases, such as cancer development. Results: We produced 100 holdouts to test the structures and diverse classification variables of diverse classifiers during the training phase. The finest performing holdouts were chosen to develop a consensus classifier and tested using a k-fold (1 ≤ k ≤5) cross-validation method. We also examined which protein properties have the biggest impact on the precise prediction of the effects of nsSNVs. Conclusion: Our Consensus Holdout Sampler outflanks other popular algorithms, and gives excellent results, highly accurate with low standard deviation. The advantage of our method emerges from using a tree of holdouts, where diverse LM/AI-based programs are sampled in diverse ways.

show abstract

Deep neural networks for phenotype prediction in rare diseases

Cited by 4 publications

References 27 publications

Unravelling inclusion body myositis using a patient‐derived fibroblast model

Unravelling inclusion body myositis using a patient‐derived fibroblast model

Prediction of Functional Effects of Protein Amino Acid Mutations

Prediction of Deleterious Single Amino Acid Polymorphisms with a Consensus Holdout Sampler

Contact Info

Product

Resources

About