Fatty Chain Acids Risk Factors in Sudden Infant Death Syndrome: A Genetic Algorithm Approach

Medicine and artificial intelligence (AI) have made great progress, since they have achieved unprecedented knowledge and explanations about how the human body works and about some diseases that seemed to have no way of preventing, diagnosing or treating or simply help make those processes more efficient. Sudden infant death syndrome (SIDS) could benefit from AI, since to date it has not been possible to clarify what actually causes it, devices to monitor vital signs have been used so far, an recently was made a predictive model to predict results from an autopsy in infants, however, further investigation is necessary to more effectively prevent. The main objective of this work is to be able to find a set of factors related to short chain fatty acids (SCFA) that could help to understand the risk of SIDS. Was used a public dataset named “Analysis of SCFA profile in infants dying of SIDS compared to infants dying of controls”, that contained SCFA values, of deceased children, labels them as SIDS death and from another cause (control). For pre-processing some variables were removed from the dataset. An analysis was performed with a feature ranking with genetic algorithm (GA) and risk analysis using the information of SCFAs and their relationship with SIDS, is presented. The median was calculated for each of the SFCA, which served to form two groups necessary to evaluate the risk difference and the risk relationship depending on the amount of acids present for each subject. As results, Octanoic acid represents a risk difference of 18% for the population with an amount less than 3.5 uM and an individual risk of 1.28 times. On the other hand, hexanoic and propinoic acids present a risk difference of less than 11% with a lesser amount of 23 and 128 uM respectively, as well as an individual risk of approximately 0.85 times. As conclusion there is 18% higher risk of developing SIDS if octanoic acid is less than 3.5 uM or 1.28 times greater risk. On the other hand, with hexanoic and propionic acid, they agree that there is an 11% lower risk of developing SIDS (of manner independent), if the values are less than 23 uM and 128 uM, respectively.

[1]  Y. Maehara,et al.  Diagnostic potential of stored dried blood spots for inborn errors of metabolism: a metabolic autopsy of medium-chain acyl-CoA dehydrogenase deficiency , 2018, Journal of Clinical Pathology.

[2]  Mohamed Ziad Dabaja,et al.  Metabolomics and Machine Learning Approaches Combined in Pursuit for More Accurate Paracoccidioidomycosis Diagnoses , 2020, mSystems.

[3]  Ghosh Sanchita,et al.  Evolutionary Algorithm Based Techniques to Handle Big Data , 2016 .

[4]  Diego Vallejo-Huanga,et al.  MARRSIDS: Monitoring Assistant to Reduce the Risk of Sudden Infant Death Syndrome , 2019, 2019 XXII Symposium on Image, Signal Processing and Artificial Vision (STSIVA).

[5]  T. Derks,et al.  Inborn Errors of Metabolism That Cause Sudden Infant Death: A Systematic Review with Implications for Population Neonatal Screening Programmes , 2016, Neonatology.

[6]  E. Behr,et al.  Cardiac Genetic Predisposition in Sudden Infant Death Syndrome. , 2018, Journal of the American College of Cardiology.

[7]  Rui Li,et al.  Histopathological distinction of non-invasive and invasive bladder cancers using machine learning approaches , 2020, BMC Medical Informatics and Decision Making.

[8]  R. Baer,et al.  Undiagnosed metabolic dysfunction and sudden infant death syndrome--a case-control study. , 2015, Paediatric and perinatal epidemiology.

[9]  Roger W. Byard,et al.  SIDS Sudden Infant and early childhood death: the past, the present and the future , 2018 .

[10]  José M. Celaya-Padilla,et al.  MRI signal and texture features for the prediction of MCI to Alzheimer's disease progression , 2014, Medical Imaging.

[11]  R. Salek,et al.  The Disruptive 4IR in the Life Sciences: Metabolomics , 2020 .

[12]  Yanhan Li,et al.  Medium‐chain acyl‐coenzyme A dehydrogenase deficiency: Six cases in the Chinese population , 2019, Pediatrics international : official journal of the Japan Pediatric Society.

[13]  M G Blitzer,et al.  Biochemical diagnosis of fatty acid oxidation disorders by metabolite analysis of postmortem liver. , 1994, Human pathology.

[14]  Jimmy D Bell,et al.  Predicting and elucidating the etiology of fatty liver disease: A machine learning modeling and validation study in the IMI DIRECT cohorts , 2020, PLoS medicine.

[15]  Francesco Falciani,et al.  GALGO: an R package for multivariate variable selection using genetic algorithms , 2006, Bioinform..

[16]  Zhijie Ren,et al.  Lilikoi V2.0: a deep learning–enabled, personalized pathway-based R package for diagnosis and prognosis predictions using metabolomics data , 2020, bioRxiv.

[17]  T. Rognum,et al.  The Sudden Infant Death Syndrome Gene: Does It Exist? , 2004, Pediatrics.

[18]  Binyam Bogale,et al.  Development of a targeted client communication intervention to women using an electronic maternal and child health registry: a qualitative study , 2020, BMC Medical Informatics and Decision Making.

[19]  C. V. Martínez,et al.  Alimentación y nutrición. Manual Teórico-Práctico. 2a Ed. , 2005 .

[20]  Samira Shamsir,et al.  Smart Infant-Monitoring System with Machine Learning Model to Detect Physiological Activities and Ambient Conditions , 2020, 2020 IEEE International Instrumentation and Measurement Technology Conference (I2MTC).

[21]  R. Knight,et al.  A Universal Gut-Microbiome-Derived Signature Predicts Cirrhosis. , 2020, Cell metabolism.

[22]  D. Gozal,et al.  Prevalence of medium-chain acyl-coenzyme A dehydrogenase deficiency in the sudden infant death syndrome. , 1993, The Journal of pediatrics.

[23]  Li-Yeh Chuang,et al.  Evaluation of Breast Cancer Susceptibility Using Improved Genetic Algorithms to Generate Genotype SNP Barcodes , 2013, IEEE/ACM Transactions on Computational Biology and Bioinformatics.

[24]  Emma Schymanski,et al.  Computational Metabolomics: From Cheminformatics to Machine Learning (Dagstuhl Seminar 20051) , 2020, Dagstuhl Reports.

[25]  Nigel J. Martin,et al.  Machine Learning Approaches to Determine Feature Importance for Predicting Infant Autopsy Outcome , 2020, medRxiv.

[26]  M. V. van Zelm,et al.  Review article: short chain fatty acids as potential therapeutic agents in human gastrointestinal and inflammatory disorders , 2018, Alimentary pharmacology & therapeutics.

[27]  N. Makita,et al.  Cardiac Ion Channel Gene Mutations in Sudden Infant Death Syndrome , 2008, Pediatric Research.