Dear Dr. Afgan,
I am in a lab that focuses on NGS analysis for Zebrafish Bis-seq dataset, and is about to use the Bismark tool (http://www.bioinformatics.babraham.ac.uk/projects/bismark) for that purpose. We were very excited to learn that Bismark works in the Galaxy project that you are the main author of, and have been trying to install Bismark into Galaxy for the past few weeks. We have tried and received the following:
1. Start up our own instance on Amazon AWS using ami-da58aab3, galaxy-cloudman-2011-03-22, and following the instructions from Galaxy Wiki to install Bismark manually, including the necessary modifications on the files "universe_wsgi.ini", "tool_conf.xml", "tool_data_table_conf.xml". However, the installed Bismark kept generating error messages that we could not fully resolve.
2. We later noticed that Bismark can be installed from Tool Shed in a rather automatic way. But in order to take advantage of that, we need to update the galaxy system into more recent version in order to have the recent/automatical Tool Shed. Oddly the link in the Cloudman Admin page for "Update Galaxy from a provided repository" doesn't work for us. So we tried to update the galaxy manually, then the Bismark from Tool Shed. It worked partially, as some tools are functional but not others, including the Bismark. We are in the process of configuring individual tools as well as Bismark.
3. During the frustration, I kept wondering whether there is a more systemic and better documented way of upgrading Galaxy and installing its tools such as Bismark. Although we may be able to figure out all the problems eventually, the time and work spent on it is rather expensive as compared to a stand-alone Linux box.
4. So in looking out for a more systemic and better documented solution, I just came across your site for the mi-deployment (https://bitbucket.org/afgane/mi-deployment). By reading its overview, it seems this is exactly the solution I have been looking for (sorry for my ignorance on this matter, as I should have tried it from the beginning).
I am in the process of trying mi-deployment. But in the meantime, I would like to ask a couple of questions:
1. Is "mi-deployment" the right track I should be following, or I am still in the wrong place for my situation? Are there more detailed instructions on doing it? Do you have any other suggestions? I consider myself and my group with fair literacy on bioinformatic tools, NGS tools, and how Linux system works (but not experts on it).
2. Per your instruction on mi-deployment, we have established a "CloudBioLinux" Ubuntu 12.04 instance (version dated to December 2012) and have set environment variables for both access key and secret access key, in addition to using pip to install boto and fabric. We've attempted to create an EBS volume and link it to our EC2 instance, but it does not show up within the "/dev" directory following a kernel restart. How may we get it to display such that we can then mount it with "sudo mount /dev/VOLUME /mnt/galaxyTools"?
3. Is it possible that I could reach you by phone for more detailed discussion?
Thanks for your great work on Galaxy, and I look forward to hearing from you.
Chun-Yuan
--
XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX
Chun-Yuan Huang, Ph.D.
Research Scientist
Huangc at uwm.edu
chunyuan.huang at gmail.com
Laboratory for Public Health Informatics and Genomics
Lapham 234, Zilber School of Public Health
University of Wisconsin – Milwaukee
XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX