search

liulab-dfci / MAESTRO

Single-cell Transcriptome and Regulome Analysis Pipeline
GNU General Public License v3.0
276 stars 76 forks source link

A hardcoded path of LISA in the docker image #82

Closed QiangTU closed 3 years ago

QiangTU commented 3 years ago

Hello,

I am using MAESTRO docker image winterdongqing/maestro, 1.2.1, 57f41ffbe200. When I ran the scRNA-seq module, it reported errors like:

unable to open file: name = '/project/dev/qqin/LISA/lisa_web/cistromedb_data/hg38/cluster_human/DNase_median_for_each_cluster.h5'

This error should be caused by a hardcoded path of LISA. I changed to web LISA, then this module finished successfully.

crazyhottommy commented 3 years ago

Thanks for reporting this! @DongqingSun can you please fix this?

DongqingSun96 commented 3 years ago

Hi @QiangTU,

The error was caused by the hardcoded path. Please download the pre-computed datasets from CistromeDB and update the configuration file following the instructions shown in LISA README.

If it doesn't work, please feel free to let us know. Thanks!

Dongqing Sun

crazyhottommy commented 3 years ago

@QiangTU we are releasing MAESTRO1.2.1 by the end of this week which includes LISA2. LISA2 is a re-write of the old LISA and is much faster https://github.com/liulab-dfci/lisa2

QiangTU commented 3 years ago

Congratulations! But, you mean 1.3? v1.2.1 has been released in Aug. :-D

crazyhottommy commented 3 years ago

sorry, I meant to say v1.2.2 :) and in one more week, we will release multi-sample scATAC v1.3.0