Setting up BUSCO well
1
1
Entering edit mode
19 months ago
Theiser • 0

Hello fellow Biostars,

while reading the answers of you folks to difficult questions in this forum, I figured that, if anyone in the internet, only you can save a desperate man like me, who is new to bioinformatics, with your impressive knowledge and experience! I am looking forward to you answers and do not think that you guys will be intimidated by a post that is a bit longer!

So I recently installed BUSCO, and while the test busco run yields the exact log that I am supposed to yield and the analysis do run without any errors, I still do not know whether I have made any mistakes. I do not want to discard all of my analysis in the end because I made a decisive mistake in the beginning, Therefore I wanted to ask you guys a couple of questions. Specifically there are a few tips in the user guide that I chose to ignore, because I thought that I would crush my entire seemingly working installation, if I tried to implement them in the wrong manner (https://busco.ezlab.org/busco_userguide.html#mandatory-arguments).

Info: I am trying to examine the quality of genome assemblies of many plants and algae. I chose the installation via anaconda.

1. At first I wanted to use metaeuk instead of augustus, but according to the user guide, the anaconda metaeuk version does not work for the current BUSCO pipeline. So I installed it manually instead, but it does not seem to be installed as a part of busco (I can see that by typing "busco -h". Normally, there would be optional arguments for metaeuk, but there are none, so busco does not recognize it or something.). Consequently, I decided to just use augustus instead. There are a few tips regarding augustus in the user guide of BUSCO. For example, it says that I should use the 3.2.3 version and type in "-- augustus" in the command line. In Addtion, there is a section that says that I should set a few variables as the following:

export PATH="/path/to/AUGUSTUS/augustus-3.2.3/bin:$PATH" export PATH="/path/to/AUGUSTUS/augustus-3.2.3/scripts:$PATH"

export AUGUSTUS_CONFIG_PATH="/path/to/AUGUSTUS/augustus-3.2.3/config/"

I havent done either of them. I use the version 3.3.3., the analysis runs augustus prediction regardless of whether I am typing "-- augutstus" or not and I also did not type in the export commands, because they are part of the section for the Manual installation for BUSCO and I installed it as a conda package, not manually. But is it for some reason I do not know critical to do those steps anyway?

1. Moreover, there is a sentence in the README.md stating "Do not forget to edit the config/config.ini file to match your environment. The script scripts/busco_configurator.py can help with this." I have not run the script because, as I said, I do not want to break something when it is all seemingly working fine right now. I do know what an environment is, but I do not know what is meant by "paths matching my environment". I can find and edit the config file and all the paths that are declared in the file are findable on my computer. So everything should be fine I guess?

I am verry sorry for the long post, but I hope you guys can help me!

Best regards, Thies

busco Assembly gene sequence genome • 1.8k views
1
Entering edit mode

What exactly is there to help with?

You said the test run was successful. It seems that you are ignoring lots of installation recommendations for no obvious reason. I don't think anyone can troubleshoot potential problems stemming from your installation - they may not happen at all. It's like saying that you are planning to drive 95 in an area where the speed limit is 65, and asking us predict what will happen. Maybe nothing bad will happen, but with your actions the chance of something bad happening is increased.

0
Entering edit mode

OP needs to be careful, though, what a successful testrun means. With a previous BUSCO version, I once had the .odb databases unreachable due to filesystem permissions. Did that stop BUSCO from running? No, it just ran everything against the integrated mouse (I think it was mouse) database. It reported that had done so in the log file, but since the run didn't fail, I never felt the need to check the logs at the time. Long story short: a successful BUSCO run doesn't necessarily mean that it actually did what one wants it to do.

0
Entering edit mode

Exactly! A test run being "successfull" does not mean that I will be satisfied with the results. As I explained in my post, it would help me to know why exactly it is recommended to do the things that I left out, so that I can understand what problems might occure. For example, I do not see any obvious reason why running Augustus 3.3.3 instead of 3.2.3 would cause any problems. Coming back to the driving analogy: If you would tell a novice driver that they could get catched speeding or crash into a tree by driving 95 instead of 65, they would be able to make their own decision whether they really want to do it or not. ;)

Maybe you will answer that you do not know why, and that this is just the way you should do things, but even this would be better than nothing! :)

1
Entering edit mode
19 months ago
Theiser • 0

My question has been answered elsewhere. As assumed, it was totally fine to ignore these steps, thank you for your input! Unfortunately I do not know how to close a thread, so I will just leave it as it is.

0
Entering edit mode

Closing the thread is not the appropriate solution to indicate resolution of a question. Accepting one (or more answers) is. You did that without providing actual solution.

1
Entering edit mode
0
Entering edit mode

Yes I did, I asked whether my actions could lead to any errors in busco, and if yes, why. As I indicated in my last comment, the answer is that all I did was fine, since the extra steps would only have been neccessary, if I had installed busco manually and not as a conda package (as I assumed in my post).