| The protection of ionizing radiation and its damage treatment have been paid much attention in specialty medicine.In recent years,the countries with nuclear weapons have been increasing,and the number of nuclear power plants has been increasing.The risk of people suffering from ionizing radiation rises.With an aim to prevent and treat the injury caused by ionizing radiation,the research and development of new safe and effective radioprotective drugs are becoming more and more urgent.In the process of R&D of radioprotective drugs,it is significant to find lead compounds with radioprotective activity.In order to find the compounds with radioprotective activity,a great deal of researches have been carried out in the United States,Russia,China,France and other countries in the 1950 s.Generally,these researches adopt cell and animal experiment methods.For example,7,000 compounds and Chinese herbal medicines have been screened by mice in China.Due to the characteristics of cells and animal experiments,these require a high cost of manpower and material,which making it difficult to screen the compounds on a larger scale.Recently,with the rapid development of transcriptome detection technology,the cost has been decreasing,and a large number of transcriptomics data about compound stimulation have accumulated rapidly.Now,it is a new way to find the radioprotective lead compound by analying the big transcriptomics data.On the basis of these data,a large scale prediction of the radioprotective lead compound was performed in this study.Firstly,this research preprocessed the transcriptomics data in LINCS program,localized the gene expression profile data generated by the compound stimulating cells,and established the data set RRML of the reported radioprotective compound.The radioprotective lead compound was then predicted by the following three methods.(1).Prediction of the radioprotective lead compounds based on transcriptomic signatures.The first part of this method is the construction of GEPCSC and transcriptomic signatures of radiation injury.GEPCSC include the transcriptomic data of compound-stimulated cells,and the transcriptomic signatures of radiation injury reflect the cellular transcriptomics response caused by ionizing radiation.Then,based on Gene Set Enrichment Analysis(GSEA),the enrichment score reflecting the radioprotective potential of compounds was defined.Subsequently,enrichment scores were calculated for more than 30,000 compounds in the GEPCSC,and a prediction process was developed based on the enrichment scores,resulting in 287 lead compounds that stably exhibited the radioprotective potential.The compounds in RRML are significantly enriched in the predicted results,such as estradiol,estriol and genistein,indicating that the predicted results are of high reference value.Finally,this method recommended reference dose for cell experiments of the prediction of the lead compound.(2).Prediction of the radioprotective lead compound based on key gene sets.Firstly,according to the mechanism of ionizing radiation injury and the previous research on radioprotective compounds,four biofunctional topics with radioprotective effect were selected,which were antioxidant,DNA repair,promoting hematopoietic and leukocyte apoptosis negative regulation,respectively.The GO database was used to construct the key gene sets of the above 4 topics.There was a positive correlation between the gene expression in the gene set and the functional strength of the gene set.Then,the ability of more than 30,000 compounds in the LINCS program to enhance the expression of a single key gene set was evaluated by the GSEA algorithm,and the compounds were screened on the basis of this.Finally,299 lead compounds showing radioprotective potential were obtained by voting on the screening results of 4 key gene sets.The compounds in RRML are significantly enriched in the results,such as estradiol,estriol,genistein,curcumin,resveratrol and so on,indicating that the predicted lead compounds have high reference value.(3).Prediction of the radioprotective lead compound based on the similarity of transcriptomics responses.The method first selected the scoreGSEA algorithm to measure the similarity of gene expression profiles,and then calculated the similarity of gene expression profiles between more than 20,000 compounds in LINCS plan and those in RRML.Finally,according to the similarity of gene expression profiles,a prediction method for radioprotective lead compounds was developed,and 96 radioprotective lead compounds were obtained.Analysis of the results indicates that the predicted lead compound has a higher reference value.The innovation point of this article is mainly reflected in two aspects.Firstly,attempts were made to predict radioprotective compounds based on transcriptomics data.In this paper,three methods based on transcriptomics data were established to predict the radioprotective compound.Secondly,compared with the traditional cell or animal experiment methods,the data used in this study are free,and the cost is significantly reduced,the period is significantly shortened and the extension is good.What’s more,it can continue to supplement the transcriptomics data to expand the prediction range and improve the prediction quality. |
