Frequently Asked Questions (FAQs)
Your may find answer to other questions at
the Service System Discussion Board
1, Why did I fail receive the password for I-TASSER Server after my registration?
A confirmation email with password will be sent to you, after you
register your email in the I-TASSER server.
But since the confirmation emails were sent out automatically by computer, it is possible that
the confirmation email was filtered out by your email security system. So it is wise to
check your junk E-mail folders, if you did not receive a confirmation email in your inbox in time.
If you found that you did not receive the confirmation email one hour after you correctly
submitted your registration information, please contact email@example.com and let us
know the email that you registered. We can resend the registration information to you, manually.
2. Why does my password not work for I-TASSER Server?
If you copy and paste your password from the confirmation email, please remember
not to include any space before and after the password. For example, if your password is IT_8abcd, please make sure that you did not type in " IT_8abcd" or
"IT_8abcd ", which is the mistake most frequently made by new users.
3. Why does my password not work for I-TASSER Suite, I-TASSER Forum, or QUARK server?
Are you using the password obtained from I-TASSER server to download I-TASSER
Suite? The I-TASSER Server, I-TASSER Suite, I-TASSER Forum, and QUARK webserver
are independent systems. You need to register them separately for different passwords.
4. How can I predict the structure of a protein with >1,500 residues?
(Short answer) No, you can not. The I-TASSER server only accepts for sequence
below 1,500 residues.
(Long answer) The server version of I-TASSER has not been optimized for
multi-domain proteins prediction. Therefore, if you upload a large protein
with more than one domain, more often than not you will only get accurate
prediction for one of your domain. If you have a protein longer than 1,500
residues, you can split the sequence into domains either by
Pfam search, or by
ThreaDom. You can then
submit the sequence of each domain separately. This should generate models of
5. How can I run I-TASSER suite on Mac, Windows or 32-bit Linux?
(Short answer) No, you cannot. All components of I-TASSER suite were developed
and compiled on 64-bit Linux.
(Long answer) Although we do not support it, you can try to recompiled a small
number of programs in the I-TASSER suite to port to other operating system. All
source-available programs are located at "src/" directory of I-TASSER suite.
I-TASSER suite was mostly written in Fortran and Perl, as well as a small amount
of C++ and JAVA. The way it parses paths makes it very hard to port to non-Unix
system such as Windows. Also, as I-TASSER suite is memory intensive, there is
no point trying to make I-TASSER run on 32-bit system where no more than 4Gb
of memory could be addressed.
6. Why my SAXSTER prediction did not return any result?
7. How can I use I-TASSER models to study the effect of mutations on the overall topology?
(Short answer) No, you cannot. Overall topology of I-TASSER models for different
mutant of the same protein is usually very similiar.
(Long answer) A structure model can tell you where is the mutation, e.g. if it
is near ligand binding site, an active site, or at protein-protein interaction
interface, etc. However, unless you have introduced a substantial number of
mutations, or introduce long insertion/deletetion, different mutants of the
same protein usually have similar final I-TASSSER structures. This is in
accordance with our observation on experimental structure in PDB, wild type
protein and mutant protein of a small number of substitution usually shares
almost identical conformation, even if their function might differ
significantly. We have developed a pipeline called STRUM
to address the effect of single mutations on a protein, specifically the
stability differences between native protein and its mutants.
8. How do I report bugs for web servers and bioinformatics tools in Yang Zhang lab?
If you have successfully submit jobs via web server, but the webserver did not
send any result back after two weeks, you should check whether your input is
incorrect. If you are sure it is a bug, please report it to
firstname.lastname@example.org using the
email you used for job submission and report the web service you used and your
If you find a bug in standalone software downloaded from Yang Zhang lab website,
such as I-TASSER suite, please report it to the
Service System Discussion Board. You need to provide the following information:
 input files,
 output files,
 command you used to run the software and its screen output,
 if you recompile the software yourself, instead of using our precompiled
binaries, please report the compiler version (e.g. gfortran 4.8) and
compilation command you used,
 the operating system (e.g. Ubuntu Linux Trusty 14.04 amd64, or Mac OSX
10.11.1 EL Capitan) and CPU description; under Linux please send the output of:
uname -a; lsb_release -a; head -n 25 /proc/cpuinfo; ulimit -a; free
9. What are the 'top 10 templates used by I-TASSER' in the Result page?
I-TASSER modeling starts from the structure templates identified by LOMETS from
the PDB library. LOMETS is a meta-server threading approach containing multiple
where each program can generate tens of thousands of templates. I-TASSER
only uses the templates of the highest significance in the threading
alignments, which are measured by the Z-score (the difference between the
raw and average scores in the unit of standard deviation).
The top 10 templates are the 10 templates selected from the LOMETS
threading programs. Usually, one (or two) template of the highest Z-score
is selected from each threading program, where the threading programs are
sorted by the average performance in the large-scale benchmark test experiments.
10. What is the 'top 5 models predicted by I-TASSER' in the Result page?
For each target, I-TASSER simulations generate tens of thousands conformations
(called decoys). To select the final models, I-TASSER uses SPICKER program to
cluster all the decoys based on the pair-wise structure similarity, and report
up to five models which corresponds to the five largest structure clusters.
In Monte Carlo theory, the largest clusters correspond to the states of the
largest partition function (or lowest free energy) and therefore have the
highest confidence. The confidence of each model is quantitatively measured by
C-score (see below).
Since the top 5 models are ranked by the cluster size, it is possible that
the lower-rank models have a higher C-score.
Although the first model has a higher C-score and a better quality in
most cases, it is not unusual that the lower-rank models have a better quality
than the higher-rank models.
If the I-TASSER simulations converge, it is possible to have less than 5 clusters
generated. This is usually an indication that the models have a good quality
because of the converged simulations.
11. What are 'Proteins structurally close to the target in the PDB' in the Result page?
After the structure-assembly simulation,
I-TASSER use TM-align program to match the first I-TASSER
model to all structures in the PDB library. This section reports the top
10 proteins from the PDB which have the closest structural similarity
(i.e. the highest TM-score) to the
predicted I-TASSER model. Due to the structural similarity, these proteins
often have similar function to the target. However, users are encouraged to
use the data in 'Predicted function using COACH' to infer the biological
function of the target protein, since COACH has been extensively trained to
derive function from multi-source of sequence and structure features which
has on average a much higher accuracy than the function annotations
derived only from the global structure comparison.
12. What is C-score in the Result page?
C-score is a confidence score for estimating the quality of predicted models
by I-TASSER. It is calculated based on the significance of threading template
alignments and the convergence parameters of the structure assembly
simulations. C-score is typically in the range of [-5,2], where a C-score of
higher value signifies a model with a high confidence and vice-versa.
13. What is TM-score in the Result page?
TM-score is a recently proposed scale for measuring the structural similarity
between two structures (see Zhang and Skolnick, Scoring function for automated
assessment of protein structure template quality, Proteins, 2004 57: 702-710).
The purpose of proposing TM-score is to solve the problem of RMSD which
is sensitive to the local error. Because RMSD is an average distance of all
residue pairs in two structures, a local error (e.g. a misorientation of the
tail) will arise a big RMSD value although the global topology is correct.
In TM-score, however, the small distance is weighted stronger than the big
distance which makes the score insensitive to the local modeling error.
A TM-score >0.5 indicates a model of correct topology and a TM-score<0.17
means a random similarity. These cutoff does not depends on the protein length.
14. What is difference and relationship between C-score and TM-score?
TM-score (or RMSD) is a known standard for measuring structural similarity
between two structures which are usually used to measure the accuracy of
structure modeling when the native structure is known, while C-score is
a metric that I-TASSER developed to estimate the confidence of the modeling.
In case where the native structure is not known, it becomes necessary to
predict the quality of the modeling prediction, i.e. what is the distance
between the predicted model and the native structures? To answer this
question, we tried predicted the TM-score and RMSD of the predicted models
relative the native structures based on the C-score.
15. Why some lower-rank models have higher C-score in the Result page?
In a benchmark test set of 500 non-homologous proteins, we found that C-score
is highly correlated with TM-score and RMSD. Correlation coefficient of C-score
of the first model with TM-score to the native structure is 0.91, while the
coefficient of C-score with RMSD to the native structure is 0.75. These data
lay the base for the reliable prediction of the TM-score and RMSD using
C-score. In the output section, I-TASSER only reports the quality prediction
(TM-score and RMSD) for the first model, because it was found that
the correlation between C-score and TM-score is weak for lower rank models.
However, the C-score is listed for all models just for a reference.
We have found that
the cluster size is more robust than C-score for ranking the predicted
models. The final I-TASSER models are therefore ranked based on
cluster size rather than C-score in the output. Nevertheless, the
C-score has a strong correlation with the quality of the final models,
which has been used to quantitatively estimate the RMSD and TM-score
of the final models relative to the native structure. Unfortunately,
such strong correlation only occurs for the first predicted model from
the largest cluster. Thus, the C-scores of the lower-rank models
(i.e., models 2-5) are listed only for reference and a comparison
among them is not advised. In other word, even though the lower-rank
models may have a higher C-score than the first model in some cases,
the first model is on average the most reliable and should be
considered if without special reasons (e.g., from biological sense or
16. Why do I-TASSER models have worse ramachandran plot than other structure prediction program? How can I refine them using FG-MD and ModRefiner?
Different from traditional homology modelling programs that build structure
models based on a single template, I-TASSER server builds models by reassmbling
the structural fragments from multiple templates.
As a result, I-TASSER models sometime have more backbone torsion angles within
energetically unfavorable regions in the Ramachandran plot; this happens more
often for the difficult targets that have no close homologous templates.
This does NOT mean I-TASSER is worse than homology modelling (in contrast, homologous
models are almost certainly incorrect for these targets).
When I-TASSER performs simulation, it generates thousands of decoy
structures. These structures are clustered by structure similarity. The top
structure model comes from structure averaging of largest cluster. The
structure averaging improves the global topology (in terms of TM-score and
RMSD between model and native structure). But it can make the local structure
such as backbone torsion angles worse, for the hard targets in which the distribution
of the structural decoys is diverged.
At the end of the I-TASSER pipeline, FG-MD or
ModRefiner is used to refine the structure to correct the local
structures. So users do not have to further refine the structure themselves.
In general, it is very difficult for structure refinment to further improve
the local structure without worsening the global topology.