TMHBenchmark

From Rost Lab Open
Jump to: navigation, search

Contents

Intro

NOTE: THIS SERVER IS CURRENTLY OFFLINE. WE WOULD WORK TO BRING IT BACK ONLINE IF ENOUGH INTEREST IS EXPRESSED PLEASE EMAIL gyachdav at rostlab dot com FOR ANY QUESTIONS.

The TMH Benchmark Server was created to allow the community to evaluate new methods for membrane helix structure prediction using a standard set of evaluations that test many aspects of TMH prediction. This work was originally published by our group for the field of known TMH prediction methods including simple hydrophobicity scale methods and more advanced servers that use Hidden Markov Models, neural nets, etc. With the implementation of this server, one can get an idea of a new method's performance in a matter of minutes along with the ranking of the method among existing methods.

Reference

TMH Benchmark is grounded in the original work:

Chen, C.P., Kernytsky, A. and Rost, B. Transmembrane helix predictions revisited. Protein Science, 11, 2774-2791 (2002).

When referring to this work please cite:

Kernytsky, A. and Rost, B.Static benchmarking of membrane helix predictions. Nucl Acids Res, 31, xx-xx (2003). In press.

Availability

Coming soon

Sample Output

<a name="hi"></a> <b>Table 1. High Resolution Accuracy</b> (Sorting on column 1 - highlighted)

<col align="left"><col align="center" width="61"><col align="center" width="61"><col align="center" width="61"><col align="center" width="61"><col align="center" width="61"><col align="center" width="61"><col align="center" width="61"><col align="center" width="61"> <tbody></tbody>
Per-segment accuracyPer-residue accuracy
<a href="http://cubic.bioc.columbia.edu/cgi-bin/var/kernytsky/tmh/h-scale-input2.cgi?id=00513&sorthi=0&sortlo=1#hi">Method</a> <a href="javascript:open_window('Method')">?</a><a href="http://cubic.bioc.columbia.edu/cgi-bin/var/kernytsky/tmh/h-scale-input2.cgi?id=00513&sorthi=1&sortlo=1#hi">Qok</a> <a href="javascript:open_window('Qok')">?</a><a href="http://cubic.bioc.columbia.edu/cgi-bin/var/kernytsky/tmh/h-scale-input2.cgi?id=00513&sorthi=2&sortlo=1#hi">Qhtm
 %obs</a> <a href="javascript:open_window('Qhtm_obs')">?</a>
<a href="http://cubic.bioc.columbia.edu/cgi-bin/var/kernytsky/tmh/h-scale-input2.cgi?id=00513&sorthi=3&sortlo=1#hi">Qhtm
 %prd</a> <a href="javascript:open_window('Qhtm_prd')">?</a>
<a href="http://cubic.bioc.columbia.edu/cgi-bin/var/kernytsky/tmh/h-scale-input2.cgi?id=00513&sorthi=4&sortlo=1#hi">Q2</a> <a href="javascript:open_window('Q2')">?</a><a href="http://cubic.bioc.columbia.edu/cgi-bin/var/kernytsky/tmh/h-scale-input2.cgi?id=00513&sorthi=5&sortlo=1#hi">Q2T
 %obs</a> <a href="javascript:open_window('Q2T_obs')">?</a>
<a href="http://cubic.bioc.columbia.edu/cgi-bin/var/kernytsky/tmh/h-scale-input2.cgi?id=00513&sorthi=6&sortlo=1#hi">Q2T
 %prd</a> <a href="javascript:open_window('Q2T_prd')">?</a>
<a href="http://cubic.bioc.columbia.edu/cgi-bin/var/kernytsky/tmh/h-scale-input2.cgi?id=00513&sorthi=7&sortlo=1#hi">Q2N
 %obs</a> <a href="javascript:open_window('Q2N_obs')">?</a>
<a href="http://cubic.bioc.columbia.edu/cgi-bin/var/kernytsky/tmh/h-scale-input2.cgi?id=00513&sorthi=8&sortlo=1#hi">Q2N
 %prd</a> <a href="javascript:open_window('Q2N_prd')">?</a>
PHDpsihtm088499988076838680
HMMTOP28399998069898871
DAS7999967248949662
TopPred27590907764839069
TMHMM17190908068818972
SOSUI7188867566748069
PHDhtm076983817876828479
KD6594896779665267
PHDhtm086477767876828479
GES6497907174726669
PRED-TMR6184907658859466
Ben-Tal6079897253809563
Eisenberg5895896977685768
Hopp-Woods5693866280614367
WW5495917171726767
Roseman5294835883583466
Av-Cid5293836083583972
Levitt4891845980583867
A-Cid4795835880563766
Heijne4593826185583464
Bull-Breese4592825585552766
Sweet4390836383604369
Radzicka4093795685552663
Nakashima3988836084583663
Fauchere3692805684563165
Lawson3386795584542763
EM3192775785552864
Wolfenden2843626228569756
YOU000480010048
</p>

.


<a name="lo"></a> <b>Table 2. Low Resolution Accuracy</b> (Sorting on column 1 - highlighted)

<col align="left"><col align="center" width="61"><col align="center" width="61"><col align="center" width="61"><col align="center" width="61"><col align="center" width="61"><col align="center" width="61"><col align="center" width="61"><col align="center" width="61"> <tbody></tbody>
Per-segment accuracyPer-residue accuracy
<a href="http://cubic.bioc.columbia.edu/cgi-bin/var/kernytsky/tmh/h-scale-input2.cgi?id=00513&sorthi=1&sortlo=0#lo">Method</a> <a href="javascript:open_window('Method')">?</a><a href="http://cubic.bioc.columbia.edu/cgi-bin/var/kernytsky/tmh/h-scale-input2.cgi?id=00513&sorthi=1&sortlo=1#lo">Qok</a> <a href="javascript:open_window('Qok')">?</a><a href="http://cubic.bioc.columbia.edu/cgi-bin/var/kernytsky/tmh/h-scale-input2.cgi?id=00513&sorthi=1&sortlo=2#lo">Qhtm
 %obs</a> <a href="javascript:open_window('Qhtm_obs')">?</a>
<a href="http://cubic.bioc.columbia.edu/cgi-bin/var/kernytsky/tmh/h-scale-input2.cgi?id=00513&sorthi=1&sortlo=3#lo">Qhtm
 %prd</a> <a href="javascript:open_window('Qhtm_prd')">?</a>
<a href="http://cubic.bioc.columbia.edu/cgi-bin/var/kernytsky/tmh/h-scale-input2.cgi?id=00513&sorthi=1&sortlo=4#lo">Q2</a> <a href="javascript:open_window('Q2')">?</a><a href="http://cubic.bioc.columbia.edu/cgi-bin/var/kernytsky/tmh/h-scale-input2.cgi?id=00513&sorthi=1&sortlo=5#lo">Q2T
 %obs</a> <a href="javascript:open_window('Q2T_obs')">?</a>
<a href="http://cubic.bioc.columbia.edu/cgi-bin/var/kernytsky/tmh/h-scale-input2.cgi?id=00513&sorthi=1&sortlo=6#lo">Q2T
 %prd</a> <a href="javascript:open_window('Q2T_prd')">?</a>
<a href="http://cubic.bioc.columbia.edu/cgi-bin/var/kernytsky/tmh/h-scale-input2.cgi?id=00513&sorthi=1&sortlo=7#lo">Q2N
 %obs</a> <a href="javascript:open_window('Q2N_obs')">?</a>
<a href="http://cubic.bioc.columbia.edu/cgi-bin/var/kernytsky/tmh/h-scale-input2.cgi?id=00513&sorthi=1&sortlo=8#lo">Q2N
 %prd</a> <a href="javascript:open_window('Q2N_prd')">?</a>
TMHMM17291929083809192
PHDpsiHtm086795948987779296
HMMTOP26694939085839191
PRED-TMR5892939078869489
PHDhtm085786868783759094
PHDhtm075685868783759094
SOSUI4988868879728890
TopPred24884798874719389
DAS3993818665859784
Ben-Tal3579908767839585
Wolfenden2956828047769779
WW2790758183597789
GES2393687887537291
Eisenberg2090637289476391
KD1388596391425091
Sweet1187595888384487
Hopp-Woods1187585490363888
Heijne1189555191353389
Av-Cid1087585389363887
Roseman989564891343088
Nakashima988565090353487
Levitt988564991353288
Lawson886574389322483
A-Cid887575289353687
Radzicka687564191322185
Bull-Breese686564091322083
Fauchere587564391332386
EM589564191322185
YOU200710010071

.


<a name="glob"></a>

Table 3. Confusion with Globular Proteins (Sorting on column 1 - highlighted)

<col align="left"><col align="center" width="61"><col align="center" width="61"><col align="center" width="61"><tbody></tbody>
<a href="http://cubic.bioc.columbia.edu/cgi-bin/var/kernytsky/tmh/h-scale-input2.cgi?id=00513&sorthi=1&sortlo=1&sortglob=0#glob">Method</a><a href="http://cubic.bioc.columbia.edu/cgi-bin/var/kernytsky/tmh/h-scale-input2.cgi?id=00513&sorthi=1&sortlo=1&sortglob=1#glob">False Positives</a><a href="http://cubic.bioc.columbia.edu/cgi-bin/var/kernytsky/tmh/h-scale-input2.cgi?id=00513&sorthi=1&sortlo=1&sortglob=2#glob">High-resolution</a><a href="http://cubic.bioc.columbia.edu/cgi-bin/var/kernytsky/tmh/h-scale-input2.cgi?id=00513&sorthi=1&sortlo=1&sortglob=3#glob">Low-resolution</a>
YOU000
SOSUI184
TMHMM1184
PHDhtm0821923
PHDpsihtm238
Wolfenden23913
Ben-Tal3114
PHDhtm0731416
PRED-TMR481
HMMTOP2601
TopPred210811
DAS1600
WW3200
GES5300
Eisenberg6600
KD8100
Sweet8400
Hopp-Woods8900
Nakashima9000
Heijne9200
Levitt9300
A-Cid9500
Av-Cid9500
Roseman9500
Lawson9800
FM9900
Fauchere9900
Bull-Breese10000
Radzicka10000

.


<a name="sigp"></a>

Table 4. Confusion with Signal Peptides

<col align="left"><col align="center" width="100"><tbody></tbody>
Method% of proteins with signal proteins
YOU0
PHDpsihtm0823
PHDhtm0824
Wolfenden26
TMHMM134
PRED-TMR41
PHDhtm0745
HMMTOP248
Ben-Tal57
SOSUI61
TopPred282
WW90
DAS97
GES98
A-Cid99
Av-Cid99
Bull-Breese99
EM99
Eisenberg99
Fauchere99
Heijne99
Hopp-Woods99
KD99
Lawson99
Levitt99
Nakashima99
Radzicka99
Roseman99
Sweet99

.



Author

Andrew Kernytsky & Burkhard Rost

Personal tools