Identification of Superior Biochemical Markers Linked to Salinity Tolerance in Wheat Using Data Mining

The recognition of pathways involved in salt tolerance mechanisms is one of the interesting topics in plant sciences and the new data-mining techniques provide a new insight for researchers. In this research, various attribute weighting and decision tree (DT) algorithms were executed to discover biochemical markers linked to salinity tolerance. In this regard, to assess some biochemical markers such as protein, superoxide dismutase, peroxidase, catalase, ascorbate peroxidase, proline, sodium and potassium in wheat and its wild relative Aegilops crassa, a factorial experiment was conducted in a completely randomized design with three replications. The factors in the study were genotypes (Arg (salt tolerance) and Alamout (salt sensitive), and a wild relative (Ae. crassa)) and salinity (sodium chloride 0 mM and 150 mM). According to these approaches, proline, superoxide dismutase and ascorbate peroxidase can be used as biochemical markers to screen wheat genotypes for salt tolerance. Feature selection and decision tree results showed that hierarchy combination of proline, superoxide dismutase and ascorbate peroxidase can be used as biochemical markers for selection of salinity tolerant wheat. The DT Parallel Gain Ratio model when run on the Info Gain Ratiodataset and the DT Gain Ratiomodel when run on the Ruledataset were the best model in distinguishing sensitive and tolerant genotypes with 97.5% and 91.67% performances, respectively. Overall, the results showed that combining bioinformatics, and laboratory studies can lead to identifying pathways associated with salinity and provide a better understanding of the mechanisms underpinning stress tolerance.