Request for installation of ForeSeqs on rackham

[bclaremar]

This ocassion I’d like to ask for some assistance regarding the installation of ForeSeqs on uppmax for sequence imputation:

https://github.com/ddarriba/ForeSeqs/releases

It seems like a requirement for this tool is the installation of the Phylogenetic Likelihood Library (which they seemed to provide as a tarball in the guide, but it looks like that no longer works), so it would probably need to be installed following the instructions from:

https://github.com/xflouris/libpll

Thank you in advance.
Kind regards, Alan

[bclaremar]

Hi Alan,

Sorry for the delay. Perhaps you solved it yourself?
Still need support?

The instructions you provide (haven't tested) could work if you load a GCC module.

Check what is available with
ml avail gcc

Perhaps do not pick a too new one.

Good luck!
Get back if you get problems

/Björn

[VladimirAlan97]

Hi! It's been a while. Thank you for setting this up so we can still communicate! 😄

Sadly I still have not been able to install it by myself and need your assistance.

As I mentioned 3 months ago, the installation of the Phylogenetic Likelihood Library (needed to install this tool Foreseqs) is not very straightforward because the tarball for the "easy way" of installing it is not available anymore.
The "hard way" needs to be installed following the instructions from:

https://github.com/xflouris/libpll

It mentions that the running of the makefile for the installation is needed to be done either as root or using sudo, which I can't do.

Kind regards,

Alan

[VladimirAlan97]

Hi again! I was looking at the repo and just realized that I was (partially) wrong. It seems like the tarball for the library dependency (Phylogenetic Likelihood Library) is still available as from an older release (1.0.3) here: https://github.com/ddarriba/ForeSeqs/releases/download/v1.0.3/libpll-1.0.11.tar.gz

This means that it can still be installed the "easy" way (following https://github.com/ddarriba/ForeSeqs/releases/tag/v1.0.4). However it seems like it still needs to be done by someone with root access (the make install of the libpll library).

I also tried to create an image using Richel's guide (https://github.com/richelbilderbeek/create_dnabert2_with_triton_singularity_container/blob/master/dnabert2_with_triton.def), ChatGPT and my intuition I created a .def file (attached singularity_build_foreseq_def.txt), but I was only able to get it as far as the make install of foreseq (line 41 of 50) and then there was an error I couldn't solve (attached singularity_build_foreseq_error.txt).

[bclaremar]

I'll test and be back.

Richel and I don't have sudo access neither so singularity may be the way as you say.
From what I read of the instructions I did not spot the sudo line however.
I'll do my best to find all needed libraries that we already have installed in some way...

[bclaremar]

did you set prefix with ./configure --prefix=<some personal folder>

Still compiling so don't know yet if this is the final solution...

[bclaremar]

libpll is compiled and installed

continuing after some coffee with the ForeSeqs...

[bclaremar]

Foreseqs configuration DO NOT WANT to find the pll library even with my stated paths and flags both on commandline and in configuration file

troubleshooting but need to focus on other matters now

[bclaremar]

Have you thought about using the binary given in the v1.0.2 version?

[bclaremar]

That one works out of the box

[VladimirAlan97]

I'll give it a shot. I'll be using that version for my initial analyses (they've been stuck for a while), but please can you keep me updated if you are able to install any of the newer versions whenever you have time? It would be very much appreciated.

I don't know if you tried the singularity way. I did try because, as per the oficial libpll instructions (https://github.com/xflouris/libpll/) it seems you need root:

git clone https://github.com/xflouris/libpll.git
cd libpll
./autogen.sh
./configure
make
make install    # as root, otherwise run: sudo make install

However I got stuck the same in singularity. I was able to success fully install the libpll library but it couldn't be found by ForeSeqs afterwards... (I think that's whats happening in the error I attached previously).

[VladimirAlan97]

@bclaremar I've just done it and it seems to work ok though there is no man page. I'd like to hear from you if there is any update that can make any of the newer versions run properly.

In case that it's not possible to install versions 1.0.3 or 1.0.4 and you are no longer working on this issue, I'd also like to know so that I start creating my pipelines taking 1.0.2 in consideration exclusively. Thanks.

[bclaremar]

I will not go further with the non-container solution.
I can ask another colleague with container expertise. Richel has time Fri afternoon

[bclaremar]

But may I first ask you to reply on our existing email conversation? It may work now.
Please make a note here that you sent it also so that I know whether it worked or not.

[VladimirAlan97]

Hey! I don't know if you meant for me to reply to this email or the original Uppmax Support thread. I tried to reply to that thread but I'm still getting a rejected message from my mail delivery system. Alan

…

________________________________ From: Björn Claremar ***@***.***> Sent: Thursday, September 19, 2024 01:35 AM To: UPPMAX/ticket_295686 ***@***.***> Cc: ALAN VLADIMIR GODINEZ PLASCENCIA ***@***.***>; Comment ***@***.***> Subject: Re: [UPPMAX/ticket_295686] Request for installation of ForeSeqs on rackham (Issue #1) But may I first ask you to reply on our existing email conversation? It may work now. Please make a note here that you sent it also so that I know whether it worked or not. — Reply to this email directly, view it on GitHub<#1 (comment)>, or unsubscribe<https://github.com/notifications/unsubscribe-auth/ATED62YSUMRHIVY6ZT3ZZGTZXJ5DRAVCNFSM6AAAAABOBCHS2GVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDGNRQGIYTGOJXGM>. You are receiving this because you commented.Message ID: ***@***.***>

[bclaremar]

It seems that your adress still is blocked. Pity.

[VladimirAlan97]

Richel gave us some more information as to why are we still blocked. It seems like it's a very deep problem that may take a while :/