tmhmm

Branch
master
develop

R package for TMHMM [1, 2], to predict transmembrane helices in proteins.

Please note that this page is intended for academic users only. Other users are requested to contact the Software Package Manager at [email protected]

• YouTube video or download video

Install

usethis::install_github("richelbilderbeek/tmhmm")

Install TMHMM to a default folder:

library(tmhmm)
install_tmhmm("https://services.healthtech.dtu.dk/download/28c408dc-ef5e-47ad-a284-66754bcd27f7")

The URL can be obtained by requesting a download link at the TMHMM website at https://services.healthtech.dtu.dk/service.php?TMHMM-2.0. As this URL expires after four hours, tmhmm cannot do this for you.

Example

library(tmhmm)

# Use an example proteome
fasta_filename <- system.file(
  "extdata",
  "UP000464024.fasta",
  package = "tmhmm"
)

# Predict the topology
topology <- predict_topology(fasta_filename)

# Simplify the protein names
topology$name <- stringr::str_match(
  string = topology$name,
  pattern = "..\\|.*\\|(.*)_SARS2"
)[,2]

# Plot the topology
plot_topology(topology)

Each element of the topology sequence denotes the location of each amino acid:

• i: inside
• o: outside
• M: membrane

FAQ

Under which operating systems does tmhmm work?

tmhmm can only work on the set of operating systems TMHMM works on. Currently, only Linux is supported:

Operating system	Supported by TMHMM	Supported by tmhmm
AIX	Y	?
IRIX32	Y	?
IRIX64	Y	?
Linux	Y	Y
MacOS	N	N
OSF1	Y	?
SunOS	Y	?
Windows	N	N

Are there similar packages?

These are the ones I use:

• pureseqtmr: predict membrane protein topology at two levels (TMH or non-TMH),
  FOSS (i.e. no registration forms needed, thus easy install)

Note that running script shows that TMHMM is approximately 100x faster than PureseqTM:

fasta_filename <- system.file("extdata", "UP000464024.fasta", package = "pureseqtmr")
system.time({tmhmm::predict_topology(fasta_filename = fasta_filename)})
system.time({pureseqtmr::predict_topology(fasta_filename = fasta_filename)})

Results:

> system.time({tmhmm::predict_topology(fasta_filename = fasta_filename)})
   user  system elapsed 
  1.723   0.008   1.733 
> system.time({pureseqtmr::predict_topology(fasta_filename = fasta_filename)})
   user  system elapsed 
117.511   1.746  85.748 

Also when running these tools from scripts, the results are similar:

time ./bin/decodeanhmm.Linux_x86_64 lib/TMHMM2.0.model UP000464024.fasta -f lib/TMHMM2.0.options
[...]
real	0m1.379s
user	0m1.375s
sys	0m0.004s

time ./PureseqTM_proteome.sh -i UP000464024.fasta 
71
real	0m53.471s
user	1m36.354s
sys	0m1.426s

References

• [1] A. Krogh, B. Larsson, G. von Heijne, and E. L. L. Sonnhammer. Predicting transmembrane protein topology with a hidden Markov model: Application to complete genomes. Journal of Molecular Biology, 305(3):567-580, January 2001.

• [2] E. L.L. Sonnhammer, G. von Heijne, and A. Krogh. A hidden Markov model for predicting transmembrane helices in protein sequences. In J. Glasgow, T. Littlejohn, F. Major, R. Lathrop, D. Sankoff, and C. Sensen, editors, Proceedings of the Sixth International Conference on Intelligent Systems for Molecular Biology, pages 175-182, Menlo Park, CA, 1998. AAAI Press.