failing to install rMATS

Question

[thomas.heigl.ibk]

Hey Biostar Community.

Recently, I started diving into alternative splicing projects and tested DEXSeq, which works for me. A different approach would be to use rMATS, I thought, however, not an easy one as I fail to install this tool.

I ran from one error into the next and now I don't know how to proceed.

What I did was to setup an environment with conda and install the dependencies with a yml file:

name: rmats
channels:
  - conda-forge
  - defaults
dependencies:
  - python=3.6.12
  - cython=0.29.21
  - blas
  - lapack
  - gsl=2.5
  - gcc_linux-64>=5.4.0
  - gfortran_linux-64
  - cmake=3.15.4
  - samtools   
  - star       
  - pairadise 

Thereafter, I tried to build rMATS and ran into some errors, for which I needed to install other things (cannot remember anymore what exactly). And now, I got stuck.

When I build_rmats it stops at following step:

...
cd rMATS_pipeline; python setup.py build_ext;
running build_ext
cp `find ./rMATS_pipeline/build | grep so` .;
Cloning into 'PAIRADISE'...
remote: Enumerating objects: 137, done.
remote: Counting objects: 100% (83/83), done.
remote: Compressing objects: 100% (67/67), done.
fatal: the remote end hung up unexpectedlyB | 114.00 KiB/s 
fatal: early EOF
fatal: index-pack failed
(rmats) 

I repeated with new environments. Tried different approaches but somehow I always end up at this step failing.

Any suggestions? And also any suggestions what to use instead of rMATS that could be used purely inside of R, as I do alignment and counting with Rsburead and featurecounts.

[truong128]

Hello Biostar Community,

I use MacOS and I'd like to install rMATs into my Mac. I tried to install it several times and failed!!! Could you please help if you could install rMATs on Mac? Many thanks

Ryan

Answer 1

[kutscherae]

rMATS is available as a bioconda package which might work for you:

conda install -c conda-forge -c bioconda rmats=4.2.0

From this post it seems like maybe the error is from git: https://stackoverflow.com/questions/21277806/fatal-early-eof-fatal-index-pack-failed

And it looks like it's happening at this part of the rMATS build where it tries to install PAIRADISE from github: https://github.com/Xinglab/rmats-turbo/blob/v4.2.0/build_rmats#L77

PAIRADISE is only used if rMATS is run with --paired-stats. You can tell the build to skip pairadise with: ./build_rmats --no-paired-model

If you do want to use --paired-stats then you could try creating a new conda environment with these dependencies and then running ./build_rmats. Potentially you'll need to set some envionrment variables as well:

CPATH to avoid an error about zlib.h and FC to avoid an error about f77

requirements.txt:

Cython=3.0.*
cmake=3.27.*
gcc=13.*
gfortran=13.*
gsl=2.7
gxx=13.*
liblapack=3.9.*
python=3.10.*
r-base=4.3.*
r-doparallel=1.0.*
r-foreach=1.5.*
r-iterators=1.0.*
r-nloptr=2.0.*
zlib=1.2.*

conda create --prefix ./rmats_conda_env
conda activate ./rmats_conda_env
conda install -c conda-forge -c bioconda --file requirements.txt
export CPATH="${CPATH}:/path/to/rmats_conda_env/include"
export FC="$(which gfortran)"
./build_rmats

[thomas.heigl.ibk]

thanks a lot.

downloading it from bioconda worked for me. I guess all the other part belong to rMATS such as producing sashimi plots are also found there.

[Stefano]

The last post brought me far, but not yet at the finish line.

I followed exactly kutscherae's advice, but R is getting in the way:

[...]
* DONE (tibble)
ERROR: dependencies ‘MASS’, ‘mgcv’ are not available for package ‘ggplot2’
* removing ‘/home/manz/Stuff/rmats_conda_env/lib/R/library/ggplot2’

The downloaded source packages are in
    ‘/tmp/RtmpA68o69/downloaded_packages’
Updating HTML index of packages in '.Library'
Making 'packages.html' ... done
Loading required package: ggplot2
Error: could not install ggplot2
In addition: Warning messages:
1: In library(package, lib.loc = lib.loc, character.only = TRUE, logical.return = TRUE,  :
  there is no package called ‘doSNOW’
2: In library(package, lib.loc = lib.loc, character.only = TRUE, logical.return = TRUE,  :
  there is no package called ‘getopt’
3: In library(package, lib.loc = lib.loc, character.only = TRUE, logical.return = TRUE,  :
  there is no package called ‘ggplot2’
4: In install.packages(package$name, repos = repos) :
  installation of package ‘mgcv’ had non-zero exit status
5: In install.packages(package$name, repos = repos) :
  installation of package ‘ggplot2’ had non-zero exit status
6: In library(package, lib.loc = lib.loc, character.only = TRUE, logical.return = TRUE,  :
  there is no package called ‘ggplot2’
Execution halted

I'm running linux mint

uname -a:
6.8.0-38-generic #38-Ubuntu SMP PREEMPT_DYNAMIC Fri Jun  7 15:25:01 UTC 2024 x86_64 x86_64 x86_64 GNU/Linux

CPU:

 Intel(R) Pentium(R) CPU        P6000  @ 1.87GHz
flags       : fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush dts acpi mmx fxsr sse sse2 ht tm pbe syscall nx rdtscp lm constant_tsc arch_perfmon pebs bts rep_good nopl xtopology nonstop_tsc cpuid aperfmperf pni dtes64 monitor ds_cpl est tm2 ssse3 cx16 xtpr pdcm popcnt lahf_lm pti ssbd ibrs ibpb stibp dtherm arat flush_l1d

RAM: 4GB

[Jeremy Leipzig]

install ggplot2 in your conda env

conda install conda-forge::r-ggplot2