Current biology especially genomics is no-doubt a big-data science in terms of data volume and data complexity.But it has its own unique properties comparing with other big-data fields.Classification of samples with features of very high dimension is still a major challenge in computational genomics, in which feature selection is a key task.