In the whole study we find that genomic features like mutations, variations, expression, CNV etc. can be used to predict which drug among 24 anticancerous drugs, can be prefered over others. Although our genomic models performed better than models in CCLE study, the overall performance is still low. In addition to the performances, the number of required genes are also very high because different set of required genes for every genes. The other major limitation of this web server is that the study is based on genomic data of cancer cell lines which might differ from the genomics of in vivo patient data of real world. Being first of its kind, we provide a primitive form of cancer drug prioritization prediction server, which may be improved in future as per availability of patient genomics data.
| |
