You are here

Questions regarding fragmentation using Robetta

2 posts / 0 new
Last post
Questions regarding fragmentation using Robetta
#1

First I want to make sure that I understand this correctly. Fragmentation creates L-K+1 fragments, blasts them against some database and finds the homologous secondary structures that are shown below (small portion as an example).  Correct?

I actually mean to ask about the many information below. The first column shows the pdb ID, then chain, residue number, residue, then what is H, L and E? Then coordinates, then I'm not sure what the rest of the columns are. I also want to ask how I can identify the fragments that correspond to a certain fragment in my sequence, because there seems to be no information about that in the fragmentation file. Finally, Robetta always gives 9mers and 3mers, why is that, and can I generate fragments of other lengths as well?

 1n40 A   165 S H  -66.309  -17.535  177.170    3.855    9.042    22.902 7     0.000 P433 F197
 1n40 A   166 I H -104.711  -44.261 -172.345    3.855    9.042    22.902 7     0.000 P433 F197
 1n40 A   167 A H  -62.841  -30.837 -178.413    3.855    9.042    22.902 7     0.000 P433 F197
 1n40 A   168 F L -105.367   22.303  167.891    3.855    9.042    22.902 7     0.000 P433 F197
 1n40 A   169 M E  -65.784  151.772  172.783    3.855    9.042    22.902 7     0.000 P433 F197
 1n40 A   170 S L -119.403   91.178 -165.517    3.855    9.042    22.902 7     0.000 P433 F197
 1n40 A   171 S L -114.587  153.248  172.822    3.855    9.042    22.902 7     0.000 P433 F197
 1n40 A   172 A L  -68.005  -24.031 -172.183    3.855    9.042    22.902 7     0.000 P433 F197
 1n40 A   173 D L -128.425  155.523  173.971    3.855    9.042    22.902 7     0.000 P433 F197

 1yht A    71 A H  -60.918  -32.578  179.256    3.865    9.717    30.230 7     0.000 P433 F198
 1yht A    72 E H  -65.272  -27.117 -178.195    3.865    9.717    30.230 7     0.000 P433 F198
 1yht A    73 N H  -87.412  -14.097 -169.362    3.865    9.717    30.230 7     0.000 P433 F198
 1yht A    74 A L  -93.084  161.236  164.697    3.865    9.717    30.230 7     0.000 P433 F198
 1yht A    75 V E  -90.702  124.401 -178.577    3.865    9.717    30.230 7     0.000 P433 F198
 1yht A    76 Q E  -90.968  130.726  179.279    3.865    9.717    30.230 7     0.000 P433 F198
 1yht A    77 G L  -81.024  171.512  177.497    3.865    9.717    30.230 7     0.000 P433 F198
 1yht A    78 K L  -65.057  -24.081  176.723    3.865    9.717    30.230 7     0.000 P433 F198
 1yht A    79 D L  -81.104    0.318  177.046    3.865    9.717    30.230 7     0.000 P433 F198

 1m6i A    88 P H  -53.890  -23.727 -179.641    3.869    9.371    49.107 7     0.000 P433 F199
 1m6i A    89 S H  -76.655   -9.239  179.912    3.869    9.371    49.107 7     0.000 P433 F199
 1m6i A    90 F H  -81.070  -15.028  179.415    3.869    9.371    49.107 7     0.000 P433 F199
 1m6i A    91 Y L  -92.365  154.958  177.727    3.869    9.371    49.107 7     0.000 P433 F199
 1m6i A    92 V E -106.044  162.942 -179.911    3.869    9.371    49.107 7     0.000 P433 F199
 1m6i A    93 S L  -73.799  144.915  180.031    3.869    9.371    49.107 7     0.000 P433 F199
 1m6i A    94 A L  -58.576  -34.034 -179.515    3.869    9.371    49.107 7     0.000 P433 F199
 1m6i A    95 Q L  -65.783  -35.614 -179.785    3.869    9.371    49.107 7     0.000 P433 F199
 1m6i A    96 D L  -88.457  -22.462 -177.905    3.869    9.371    49.107 7     0.000 P433 F199

 

Post Situation: 
Thu, 2018-06-28 13:24
ahmadkhalifa

H/L/E are secondary structure assignments - these are DSSP letters. helix loop sheet.

The "coordinates" are backbone torsions - phi, psi, omega probably - the third is definitely omega.  Only those three are stored because fragments assume ideal coordinates (so bond lengths and angles are assumed, and sidechains are irrelevant because we repack anyway).  I don't know what the rest of the columns do without reading the code.

I think 3 and 9 was chosen empirically: 15 years ago they tried several lengths and found that ab initio performed best at those lengths, so we standardized on it.

(I am aware this is an incomplete answer, but this is what I know off the top of my head)

Thu, 2018-06-28 13:34
smlewis