BIOINFORMATICS APPLICATIONS NOTE
Vol. 28 no. 12 2012, pages 1663–1664
RFMapp: ribosome flow model application
Hadas Zur1and Tamir Tuller2,∗
1The Blavatnik School of Computer Science, Faculty of Exact Sciences and2Department of Biomedical Engineering,
Faculty of Engineering, Tel Aviv University, Ramat Aviv 69978, Israel
Associate Editor: Trey Ideker
Advanced Access publication April 11, 2012
Summary: The RFMapp is a graphical user interface application
based on the RFM (ribosome flow model), enabling the estimation
of the translation elongation rates of messenger ribonucleic acids
(mRNAs) and the profile of ribosomal densities along the mRNAs, in a
computationally efficient way. The RFMapp is based on the approach
previously described by Reuveni et al., and unlike other traditional
approaches in the field, which are mainly related to the genes’
mean codon translation efficiency, the RFM additionally considers
the codon order, the ribosomes’ size and their order. Thus, it has been
shown that RFM outperforms traditional predictors when analyzing
both heterologous and endogenous genes.
Availability and implementation:
application and guideline are available for download at:
Received on February 15, 2012; revised on February 15, 2012;
accepted on April 2, 2012
Gene translation is a complex process through which a messenger
ribonucleic acid (mRNA) sequence is decoded by the ribosome to
produce a specific protein. The elongation step of this process is
an iterative procedure in which each codon in the mRNA sequence
is recognized by a specific transfer RNA (tRNA), which adds one
additional amino acid to the growing peptide. As gene translation
is a central process in all living organisms, its understanding has
important ramifications to every biomedical field, namely human
health (Kimchi-Sarfaty et al., 2007), biotechnology (Gustafsson
et al., 2004) and evolution (Tuller et al., 2011).
In recent years, there has been a sharp increase in emerging
technologies for measuring different features related to the process
of gene translation (Ingolia et al., 2009; Schwanhausser et al.,
2011). However, this process is still enigmatic and contradicting
conclusions regarding the essential parameters that determine
translation rates appear in different studies (Plotkin and Kudla,
2010); suggesting that the rate-limiting parameters vary across
organisms, under diverse conditions and in different genes.
Ribosome flow model (RFM) is a new approach for modeling
the process of translation elongation, affording the first large-scale
analysis of gene translation elongation, while taking into account
the stochastic nature of the translation process. RFM is aimed at
∗To whom correspondence should be addressed.
capturing the effect of codon order and composition on translation
rates, the interactions between ribosomes and the characteristics
of the translation elongation process on all its various physical
The main advantage of the RFM is 2-fold. First, it considers
widely used predictors of elongation that are based on the coding
sequence, such as the codon adaptation index (Sharp and Li, 1987)
and the tRNA adaptation index (dos Reis et al., 2004).
comprehensive models of translation elongation, such as the totally
asymmetric simple exclusion process (TASEP; first suggested by
Heinrich and Rapoport, 1980), whereas its predictions are relatively
similar to the ones obtained by the TASEP.
RFM is a simple, physically plausible computational model that
is solely based on the coding sequence (i.e. a vector of codons
in each gene) and on an additional parameter (initiation rate).
The model allows for a computationally efficient analysis of the
translation process on a genome-wide scale and across all species.
Focusing on the coding sequence, we by no means to imply that
it is the only factor taking place in the determination of translation
rates. Nevertheless, as it has been widely recognized as a prime
factor in the translation elongation process, we chose to concentrate
Herein, we present RFMapp, a distributable cross-platform,
graphical user interface application based on the approach of
Reuveni et al. (2011). The new tool will serve the community
by enabling the prediction of fundamental features of the gene
translation process, including translation rates, ribosomal densities
and the relationship between all these variables. Protein abundance
levels are proportional to the multiplication of the predicted
translation rates and mRNAlevels and can be extrapolated if mRNA
levels are available (Reuveni et al., 2011). We provide below a
high-level description of RFMapp, with a small-scale example.
The usage of RFMapp is straightforward. The input is genes’
coding sequences, codon translation times, initiation rate (which
can either be global/identical for all the organism’s genes or
specific per gene) and number of approximated translation sites
(chunk size), for a certain organism. The output is a text file
of the predicted translation features and an optional graphical-
predicted ribosomal density profile. RFMapp provides the genomic
data (coding sequences and codon translation times) for six model
organisms (for a list of supported organisms, see the RFMapp
© The Author 2012. Published by Oxford University Press. All rights reserved. For Permissions, please email: email@example.com
at TEL AVIV UNIVERSITY on August 1, 2012
H.Zur and T.Tuller Download full-text
Fig. 1. (A) An illustration of the RFM model. (B) An example ribosomal
density profile of the ribosomal protein RP51B (YDR447C)—an output of
User Guide) and supports both organisms across all the domains
of life and synthetic simulations.
The RFM has two free parameters: the initiation rate λ and the
number of codons C at each ‘site’ (can be proportional to the
ribosome size). Each site has a corresponding transition rate λ, that
is estimated based on the co-adaptation between the codons of the
site and the tRNA pool of the organism [see the Methods section
of Reuveni et al. (2011)]. The output of the model consists of the
steady-state occupancy probabilities of ribosomes at each site and
As aforementioned, the input to the RFMapp includes a coding
sequence, estimated translation time for each codon, initiation rate
(λ) and chunk size (C). The mRNA molecules are coarse grained
into sites of codons according to the chunk size and the coding
sequence (in Fig. 1A, C=3); ribosomes arrive at the first site with
by another ribosome. The rate of each chunk is computed by the
RFMapp based on the translation times of the organism’s codons
and the codon composition of the chunk. In this application, the
transition time of a codon can be determined by the tRNA pool of
the organism. Briefly, taking into account the affinity between tRNA
species and codons, the translation rate of a codon is proportional to
the abundance of the tRNAspecies that recognize it (Reuveni et al.,
By the RFM, a ribosome that occupies the ith site moves, with
rate λ, to the consecutive site provided the latter is not occupied
by another ribosome. Denoting the probability that the ith site is
occupied at time t by Pi(t), it follows that the rate of ribosome
flow into/out of the system is given by λ(1−P1(t)) and λnPn(t)
respectively. The rate of ribosome ‘flow’ from site i to site i+1
is given by λiPi(t)(1−Pi+1(t)). The RFM determines the steady-
state solution of the occupation probabilities of the equations
below and specifically the rate of protein production at steady
We describe the application of RFMapp on Saccharomyces
We select the initiation time to be 0.06 and the chunk size
10 codons (approximate eukaryote ribosome footprint). The gene
FASTA file is as follows:
A text file containing the organisms’ 61 codon translation rates
must be provided. The textual output containing the predicted
translation rate, and occupation probabilities at steady state
(ribosomal density profile), is as follows:
Translation rate = 0.000599
Occupation probabilities = 0.99002 0.82625 0.78358 0.75364
0.2011 0.26172 0.19952 0.18489 0.20107 0.19337 0.17971 0.20113
The graphical-predicted ribosome density profile is shown in
H.Z. is supported by the Edmond J Safra Bioinformatics program at
Tel Aviv University.
Conflict of Interest: none declared.
dos Reis,M., et al. (2004) Solving the riddle of codon usage preferences: a test for
translational selection. Nucleic Acids Res., 32, 5036–5044.
Gustafsson,C., et al. (2004) Codon bias and heterologous protein expression. Trends
Biotechnol., 22, 346–353.
Heinrich,R. and Rapoport,T.A. (1980) Mathematical modelling of translation of
mRNA in eucaryotes; steady state, time-dependent processes and application to
reticulocytes. J. Theor. Biol., 86, 279–313.
Ingolia,N.T., et al. (2009) Genome-wide analysis in vivo of translation with nucleotide
resolution using ribosome profiling. Science, 324, 218.
Kimchi-Sarfaty,C., et al. (2007) A “Silent” polymorphism in the MDR1 gene changes
substrate specificity. Science, 315, 525–528.
Plotkin,J.B. and Kudla,G. (2010) Synonymous but not the same: the causes and
consequences of codon bias. Nat. Rev. Genet., 12, 32–42.
Reuveni,S., et al. (2011) Genome-scale analysis of translation elongation with a
ribosome flow model. PLoS Comput. Biol., 7, e1002127.
Schwanhausser,B., et al. (2011) Global quantification of mammalian gene expression
control. Nature, 473, 337–342.
Sharp,P.M. and Li,W.H. (1987). The codon adaptation index-a measure of directional
synonymous codon usage bias, and its potential applications. Nucleic Acids Res.,
Tuller,T., et al. (2011) Association between translation efficiency and horizontal gene
transfer within microbial communities. Nucleic Acids Res., 39, 4743–4755.
at TEL AVIV UNIVERSITY on August 1, 2012