Premium
Mining gene expression omnibus metadata to unravel key genes entwined with pathogenesis and prognosis of Alzheimer’s disease: A comprehensive bioinformatics approach
Author(s) -
Sree G.N.S Hema,
Rajalekshmi Saraswathy Ganesan,
Burri Raghunath R.
Publication year - 2020
Publication title -
alzheimer's and dementia
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 6.713
H-Index - 118
eISSN - 1552-5279
pISSN - 1552-5260
DOI - 10.1002/alz.042782
Subject(s) - false discovery rate , gene , disease , computational biology , value (mathematics) , bioinformatics , biology , medicine , genetics , computer science , machine learning
Background Gene expression analysis has become a valuable resource to derive new hypotheses in exploring genetic background of neurodegenerative disorders. Dearth in break through Alzheimer’s disease research to unravel i) genetic predisposition, and ii) disease specific druggable targets, prompted us to initiate and frame this project. Method Gene Expression Omnibus (GEO) database was screened with a set of prescribed inclusion and exclusion criterion. Datasets evaluated in Homosapiens containing controls and patient samples describing the genetic source and diagnostic criteria were selected for the study. Datasets evaluated in cell lines, animals, mutational studies and methylation status were excluded. Shortlisted datasets were analysed individually to retrieve Differentially Expressed Genes (DEGs) by considering False Discovery Rate (FDR) and log Fold Change (FC) through GEO2R tool. Different permutation and combinations were tried between FDR p‐value and log FC (FDR p‐value < 0.05 and log FC >2; FDR p‐value < 0.05 and log FC >1.5; FDR p‐value < 0.05 and log FC >1; and FDR p‐value <0.01 and log FC >1) to retrieve common DEGs from 70% of the datasets satisfying any of the above criteria between control and disease groups. Protein‐Protein Interactions (PPI) of common DEGs with Literature Derived Genes (LDGs) (retrieved from NCBI) was constructed to identify potential AD specific target proteins. Result Around 32 datasets were identified from GEO database meeting the prescribed criteria. Each dataset was analysed individually and DEGs were retrieved. Amongst the different combinations applied between FDR p‐value and log FC, FDR p‐value <0.01 and log FC >1 was considered suitable and this ultimately resulted in 6 datasets which were found to satisfy the fixed criteria in up‐regulated and down‐regulated DEGs as well. Amid the 6 datasets, common DEGs were found in 4 datasets (about 70%) ‐ SLC5A3 and SERPINA3 from up‐regulated DEGs; SST, RGS4, CRYM, NPTX2, RTN3, BDNF and ENC1 from down‐regulated DEGs. PPI analysis revealed significant interactions of SERPINA3, SST, RGS4, BDNF and RTN3 with LDGs. Conclusion The above analysis identified the genetic potential of SERPINA3, SST, RGS4, BDNF and RTN3 to underpin AD pathogenesis.