Rough set method accurately predicts unknown protein class/family of Leishmania donovani membrane proteome

Leishmania donovani is the primary cause of a fatal disease visceral leishmaniasis (VL) in East Africa and in the Indian subcontinent. Human beings are the only known reservoir of L. donovani and due to the emergence and the spread of drug resistance control for this disease is become worse. Therefore, identification of novel drug target is very important to develop new drug and combat drug resistance issue. Experimental determination of target is costly and time-consuming, hence it is necessary to first identify the efficient target with the accurate mathematical method and then further go for in vitro/in vivo study. Earlier we have predicted the role of protein in term of the target with Naïve Bayes probabilistic classifier on the proteins identified in our L. donovani membrane proteomics study. This time we have used alternative and the popular method named as a Rough Set method (an important part of soft computing method relevance in many real-world applications) and tried to re-visit/validate our earlier findings of L. donovani membrane proteomics and additionally decipher the unknown class/family of membrane proteins as known one. Comparing this result with other classifiers (NB, SVM, RF, C4.5 decision tree) Rough Set method has outperformed and we found the accuracy was 89.28%. This study further validates our previous finding strongly and predicts the class/family of unknown proteins which are very important for the identification and selection toward some novel drug target (still unexplored) and ultimately move in the direction of development of effective antileishmanials.