Managing risks in drug discovery: reproducibility of published findings