Abstract: [Purpose / significance] This paper is written to address the reproducibility crisis and improve the truth of scientific research. [Method / process] First the definition of reproducibility is introduced. And then selective reporting of results is analyzed to be the main incentive, including HARKing, cognitive biases and p-hacking. Following it, the TOP guidelines for promoting transparency and open science are presented in detail. [Results / conclusion] On the basis of the TOP guidelines, four practices are deeply discussed. They are registered reports , open science, preprints and TOP factor. These practices are significantly effective and should be for wider adoption, evaluation and improvement.