About TFAM TFAM is a system to classify the function of tRNAs. TFAM can automatically and confidently:
Predict initiator tRNAs in all three domains of life. Predict the charging potential of suppressors and tRNA-like RNAs As described in the TFAM Web Server publication Tåquist et al. (2007) Nucl. Acids Res.; doi:10.1093/nar/gkm393 this webserver can also predict some bacterial lysylated isoleucine tRNAs, however this functionality is less general and will not work for some bacterial taxa.
TFAM uses information from the entire sequence, and not just the anticodon, to predict function or aminoacylation charging potential. Therefore TFAM predictions are more robust to sequencing errors and natural variability than anticodon-based predictions. TFAM works by structurally aligning and scoring query sequences against families of tRNAs the function of which are known or trusted.
New March, 2007: TFAM version 1.3 released which can automatically detect anticodons and use them to compare anticodon-based identity annotations with its whole-sequence-based classifications. This is useful if you want to find tRNAs with potential conflicts between their amino acid charging specificities and their codon reading specificities, without providing annotations in the sequence identifiers.
This webserver can also return structural alignments of tRNAs sorted by predicted tRNA identity. Please see Credits below about proper citation when using the server for this purpose.
You can download a standalone version of TFAM 1.3 for free, which you must compile on a Mac or other UNIX-like system. The standalone version provides a little extra functionality including the ability to run TFAM with your own user-defined models and the ability to exclude positions (such as the anticodon) from scoring.
You can see a sample screenshot of what the TFAM webserver output looks like.
Models used in the TFAM Web Server New from version 1.0 of this Webserver and the free standalone version of TFAM are an bacterial model 0.2 that can distinguish lysidynylated isoleucine tRNAs from elongator methionine tRNAs. This model is similar to that described in Ardell and Andersson (2006) Nucl. Acids Res. 34(3):893-904 , with some additions and changes described fully in Tåquist et al. (2007) Nucl. Acids Res.; doi:10.1093/nar/gkm393 . A proteobacterial model of lysylated isoleucine tRNAs has been added and used to correct some of the misannotated isoleucine tRNAs in earlier models.
Also, new models are provided for eukaryotic cytoplasmic and archaeal tRNAs. These have only been tested for their ability to predict initiator tRNAs. Its use to predict other types of tRNAs should be considered preliminary and experimental.
See Help for more information about the models in TFAM.
Credits The TFAM webserver was designed and written by Helena Tåquist. Additional design, modifications and development by David Ardell. Yuanyuan Cui produced and researched the extended bacterial model used from TFAM v.1.0, from data generously provided by Paul Higgs. The accuracy for initiator tRNAs of the eukaryotic model used from TFAM 1.0 was verified in Drosophila by Casey Bergman. TFAM uses covariance models provided with tRNAscan-SE. This server is maintained by Emil Lundberg of the BMC Computing Department.
Citing: If you use tRNA identity predictions from this webserver in your research please cite:
Tåquist et al. (2007). TFAM 1.0: an online tRNA function classifier. Nucl. Acids Res.; doi:10.1093/nar/gkm393
If you use structural alignments of tRNAs from this webserver please additionally cite:
Lowe, T.M. and S.R. Eddy (1997) Nucl. Acids Res. 25(5):955-964
Privacy policy We are committed to protecting data privacy for our users. We maintain reasonable standards of server security and your data cannot be accessed by anyone without access to the job ID automatically generated for your job by the server. You may delete your data and TFAM analysis from our server at any time after running your TFAM analysis. All data and output generated by our server are periodically completely deleted automatically. The only data that we retain permanently are the date and time of your job, the model you used, information about your network location and web-client, and any optional information you choose to provide when you run your job.
Disclaimer This Web Server is operated in the hope that it will be useful,
but WITHOUT ANY WARRANTY; without even the implied warranty of
MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.
See a copy of the GNU General Public License for more details.
