Gene discovery for facioscapulohumeral muscular dystrophy by machine learning techniques

Abstract

Facioscapulohumeral muscular dystrophy (FSHD) is a neuromuscular disorder that shows a preference for the facial, shoulder and upper arm muscles. FSHD affects about one in 20-400,000 people, and no effective therapeutic strategies are known to halt disease progression or reverse muscle weakness or atrophy. Many genes may be incorrectly regulated in affected muscle tissue, but the mechanisms responsible for the progressive muscle weakness remain largely unknown. Although machine learning (ML) has made significant inroads in biomedical disciplines such as cancer research, no reports have yet addressed FSHD analysis using ML techniques. This study explores a specific FSHD data set from a ML perspective. We report results showing a very promising small group of genes that clearly separates FSHD samples from healthy samples. In addition to numerical prediction figures, we show data visualizations and biological evidence illustrating the potential usefulness of these results.

Update on the Molecular Aspects and Methods Underlying the Complex Architecture of FSHD

“…Nevertheless, AI models taking DEGs from muscle tissue as an input give a high accuracy for both biceps (0.90) and deltoids (0.80) using L 1 -regularized logistic regression [ 81 ]. A similar level of accuracy (0.91, 95% CI [0.907–0.913]) was yielded using an SVM to diagnose FSHD on gene expression data from skeletal muscle biopsies, whereas a previous SVM application on the same dataset reported an impressive 0.994 accuracy [ 117 , 118 ]. Even though there is only one published experiment where FSHD is diagnosed using gene expression data from blood samples, the results are promising, with a logistic regression achieving a mean AUC ranging from 0.794 (95% CI [0.618–0.961]) to 0.883 (95% CI [0.735–1.0]) [ 103 ].…”

“… A comparison of gene-expression-based AI models with data attained from (Gonza-lez-Navarro et al, 2013 [ 117 ], Gonza-lez-Navarro et al, 2015 [ 118 ], Rahimov 2012 [ 81 ] and Signorelli 2020 [ 103 ]). ( A ) Bar plot indicating the number of DEGs found in the considered studies.…”

Assessing facial weakness in myasthenia gravis with facial recognition software and deep learning

“…Artificial intelligence (AI) applications are an area of intense medical research, particularly to improve diagnosis and for home monitoring of disease severity 9 . However, to our knowledge AI applications are relatively scarce in the field of neuromuscular disorders so far and appear to be restricted to muscle imaging and gene profiling 10–14 . Assessment of facial weakness has thus far only been carried out in facial palsies and almost exclusively on photographs 15–20 .…”

“… 9 However, to our knowledge AI applications are relatively scarce in the field of neuromuscular disorders so far and appear to be restricted to muscle imaging and gene profiling. 10 , 11 , 12 , 13 , 14 Assessment of facial weakness has thus far only been carried out in facial palsies and almost exclusively on photographs. 15 , 16 , 17 , 18 , 19 , 20 As the facial features of MG patients are distinct from healthy subjects, 21 we hypothesized that a short video recording of MG patients contains information that can be used both for diagnostic and monitoring purposes.…”