Abstract
Background: The clinical diagnosis of major depressive disorder (MDD) mainly relies on subjective assessment of depression-like behaviors and clinical examination. In the present study, we aimed to develop a novel diagnostic model for specially predicting MDD.
Methods: The human brain GSE102556 DataSet and the blood GSE98793 and GSE76826 Data Sets were downloaded from the Gene Expression Omnibus (GEO) database. We used a novel algorithm, random forest (RF) plus artificial neural network (ANN), to examine gene biomarkers and establish a diagnostic model of MDD.
Results: Through the “limma” package in the R language, 2653 differentially expressed genes (DEGs) were identified in the GSE102556 DataSet, and 1786 DEGs were identified in the GSE98793 DataSet, and a total of 100 shared DEGs. We applied GSE98793 TrainData 1 to an RF algorithm and thereby successfully selected 28 genes as biomarkers. Furthermore, 28 biomarkers were verified by GSE98793 TestData 1, and the performance of these biomarkers was found to be perfect. In addition, we further used an ANN algorithm to optimize the weight of each gene and employed GSE98793 TrainData 2 to build an ANN model through the neural net package by R language. Based on this algorithm, GSE98793 TestData 2 and independent blood GSE76826 were verified to correlate with MDD, with AUCs of 0.903 and 0.917, respectively.
Conclusion: To the best of our knowledge, this is the first time that the classifier constructed via DEG biomarkers has been used as an endophenotype for MDD clinical diagnosis. Our results may provide a new entry point for the diagnosis, treatment, outcome prediction, prognosis and recurrence of MDD.
Keywords: Major depressive disorder, biomarkers, genome-wide microarray analysis, ensemble learning, gene expression profiling, neuropsychiatric disorder.
Graphical Abstract
[http://dx.doi.org/10.1007/s00146-018-0805-0]
[http://dx.doi.org/10.1080/13607863.2015.1083945] [PMID: 26381843]
[http://dx.doi.org/10.1016/j.pharmthera.2020.107494] [PMID: 31991195]
[http://dx.doi.org/10.1186/1755-7682-3-1] [PMID: 20150988]
[http://dx.doi.org/10.1016/j.neubiorev.2016.02.011] [PMID: 26906761]
[http://dx.doi.org/10.1097/YCO.0b013e32835a5947] [PMID: 23154643]
[http://dx.doi.org/10.3109/09540261.2013.825580] [PMID: 24151803]
[http://dx.doi.org/10.1017/S1461145711001246] [PMID: 21813045]
[http://dx.doi.org/10.1016/j.psyneuen.2017.10.009] [PMID: 29055264]
[http://dx.doi.org/10.3389/fnmol.2017.00272] [PMID: 28912679]
[http://dx.doi.org/10.1097/YIC.0000000000000239] [PMID: 30300165]
[http://dx.doi.org/10.1111/pcn.12299] [PMID: 25825158]
[http://dx.doi.org/10.1016/j.biopsych.2010.03.017] [PMID: 20471630]
[http://dx.doi.org/10.1038/mp.2011.166] [PMID: 22158016]
[http://dx.doi.org/10.1021/acschemneuro.1c00087] [PMID: 34008957]
[http://dx.doi.org/10.1002/adhm.201901862] [PMID: 32627972]
[http://dx.doi.org/10.1186/1471-2105-8-25] [PMID: 17254353]
[http://dx.doi.org/10.2217/14622416.7.7.1017] [PMID: 17054412]
[http://dx.doi.org/10.1038/s41598-018-34833-6] [PMID: 30405137]
[http://dx.doi.org/10.1002/ajmg.b.30272] [PMID: 16526044]
[http://dx.doi.org/10.1093/nar/gkv007] [PMID: 25605792]
[http://dx.doi.org/10.3389/fgene.2018.00589] [PMID: 30555514]
[http://dx.doi.org/10.1371/journal.pone.0184129] [PMID: 28873455]
[http://dx.doi.org/10.1093/nar/gkw1092] [PMID: 27899662]
[http://dx.doi.org/10.1038/75556] [PMID: 10802651]
[http://dx.doi.org/10.1186/gb-2003-4-5-p3] [PMID: 12734009]
[http://dx.doi.org/10.2217/bmm.14.114] [PMID: 25731213]
[http://dx.doi.org/10.1002/hbm.24282] [PMID: 30113112]
[http://dx.doi.org/10.1186/1742-2094-9-275] [PMID: 23259598]
[http://dx.doi.org/10.3390/coatings11111273]
[http://dx.doi.org/10.1002/aisy.202000084]
[http://dx.doi.org/10.1016/j.jpsychores.2019.109796] [PMID: 31470255]
[http://dx.doi.org/10.1038/nature09267] [PMID: 20703300]
[http://dx.doi.org/10.1021/acsami.0c18470] [PMID: 33373205]
[http://dx.doi.org/10.1016/j.celrep.2020.02.007] [PMID: 32101747]
[http://dx.doi.org/10.1016/j.cell.2018.07.049] [PMID: 30173917]
[http://dx.doi.org/10.2174/1567205012666150302160145] [PMID: 25731621]
[http://dx.doi.org/10.1016/j.brainresbull.2013.06.001] [PMID: 23756188]
[http://dx.doi.org/10.1016/j.dadm.2016.04.001] [PMID: 27408937]
[http://dx.doi.org/10.1007/s00406-011-0270-y] [PMID: 22057216]
[http://dx.doi.org/10.1038/npp.2009.4] [PMID: 19295509]
[http://dx.doi.org/10.1038/ng.3808] [PMID: 28288114]
[http://dx.doi.org/10.1016/j.neuron.2018.02.013] [PMID: 29526553]
[http://dx.doi.org/10.2147/NDT.S245282] [PMID: 32158215]
[http://dx.doi.org/10.3389/fncel.2017.00237] [PMID: 28860968]
[http://dx.doi.org/10.1007/s12035-016-0102-1] [PMID: 28559998]
[http://dx.doi.org/10.1371/journal.pone.0005071] [PMID: 19343227]
[http://dx.doi.org/10.1016/j.drugalcdep.2014.05.023] [PMID: 24962325]
[http://dx.doi.org/10.3389/fnmol.2019.00022] [PMID: 30800055]
[http://dx.doi.org/10.2174/1568026621666210521144534] [PMID: 34355686]
[http://dx.doi.org/10.1016/j.biopsych.2021.01.011] [PMID: 33875230]