在之前的寫的教程中髓棋，我講到從NCBI下載SRA數(shù)據(jù)時，都會教大家配置Aspera缸沃，提高下載速度抒痒。
之前一直有人問我為什么最近無法prefetch下載數(shù)據(jù)時即便指定了ascp幌绍，但是依舊用HTTPS呢？我之前一直以為是該數(shù)據(jù)并不支持ascp的下載方式故响，結(jié)果現(xiàn)在發(fā)現(xiàn)是所有NCBI的SRA數(shù)據(jù)都不在支持了纷捞。
有人在Github上提問為什么 prefetch不能使用fasp下載，作者的回復內(nèi)容如下
NCBI has shifted to using cloud-style object stores. The Aspera algorithm requires random disk access to perform its function, and random access to object stores gives very poor performance. If Aspera comes out with a server that works well on cloud and object stores, we will support it, but currently there is no Aspera support available to my knowledge.
prefetch is not canceling ascp support, as you say, but will remove the option from the command line in the next release because there will not be an Aspera server available for AWS, GCP or NCBI, making the option useless and misleading. It will be added back in if the situation changes.
見 https://github.com/ncbi/sra-tools/issues/255#issuecomment-566068771
也就是說NCBI改變了數(shù)據(jù)存放形式被去，導致原本ascp的下載方式效率低下，于是干脆就放棄了奖唯。
怎么說呢惨缆，真的坑啊丰捷！
