SSIPe

This forum is shown on the index page along with all topics.

Moderator: robpearc

z.heidari
Posts: 2
Joined: Wed Jan 19, 2022 8:46 pm

SSIPe

Post by z.heidari »

I have tried the standalone version of SSIPe method to calculate the changes in binding energy upon mutations.
I tried to reproduce your results on your train data set. My results are slightly different than yours. I got the same EvoEF score but different SSIP scores.
I am wondering if you have any idea what could be wrong. I am also wondering if you did any optimization or minimizations on the pdb files before running SSIPe method or did you just use the original pdb files.
xiaoqiah
Posts: 6
Joined: Tue May 04, 2021 5:05 pm

Re: SSIPe

Post by xiaoqiah »

Thank you for the interest.

Could you post a few examples for the differences. The SSIP scores are calculated based on interface structural and sequence alignments. And effective number of interface sequence alignments is calculated in sequence hit-order dependent manner. I guess the difference may be caused by different sequence alignment orders.

To check this, you can run some examples using our SSIPe server and compare it with the results obtained by the standalone version. You can compare the sequence_align.out files. For an example, please refer to https://zhanggroup.org/SSIPe/example/.

Besides, there is indeed a PDB structure reparation procedure by EvoEF. But as you mentioned, you got identical values for EvoEF scores, so EvoEF minimization should not be the problem.
z.heidari
Posts: 2
Joined: Wed Jan 19, 2022 8:46 pm

Re: SSIPe

Post by z.heidari »

Thanks so much for your help. Here is the results from running SSIPe on couple of mutations on 5M2O using both website and standalone version:

from the website:
DDG SSIPscore EvoEFscore mutations
1.200 0.459 1.930 EA80A;
0.956 0.638 0.830 HA117A;
1.530 0.978 1.780 IB17A;
0.611 0.080 1.020 KA73A;
0.663 0.080 1.170 KB32A;
2.021 1.879 1.280 LA75A;
2.211 1.879 1.840 LB65A;
0.671 0.273 0.780 NA120A;
0.419 0.273 0.040 NA64A;
0.825 0.273 1.230 NA71A;
0.792 0.080 1.550 QB25A;
0.727 0.080 1.360 QB61A;
0.647 0.273 0.710 SB18A;
1.526 1.140 1.420 VB21A;

Structure-based interface MSA by iAlign from NIL [download] [Explanation]
Interface residues:
chain_ID_A: A
residue_A: 33 34 35 36 62 63 64 67 69 70 71 72 73 75 77 78 79 80 81 113 115 117 119 120 121 122 123 124
chain_ID_B: B
residue_B: 10 12 16 17 18 21 22 24 25 28 29 32 33 58 61 62 64 65 67
Alignments:
PDB chA chB ifResNum alnResNum score ifSeqA ifSeqB
target A B 47 47 1.000 SAMDMTNIGANFKLDKGEPKEHSNDGSQ NDDISVLMQANKYQQSCLL


Sequence-based interface MSA by PSI-BLAST from STRING [download] [Explanation]
Interface residues:
chain_ID_A: A
residue_A: 33 34 35 36 62 63 64 67 69 70 71 72 73 75 77 78 79 80 81 113 115 117 119 120 121 122 123 124
chain_ID_B: B
residue_B: 10 12 16 17 18 21 22 24 25 28 29 32 33 58 61 62 64 65 67
Alignments:
Name chA chB ifResNum alnResNum linkscore ifSeqA ifSeqB
target A B 47 47 1.000 SAMDMTNIGANFKLDKGEPKEHSNDGSQ NDDISVLMQANKYQQSCLL


---------------------------------------------------------------------------------------------------
from standalone run:
DDG SSIPscore EvoEFscore mutations
0.863 0.000 1.930 EA80A;
0.632 0.196 0.830 HA117A;
1.228 0.567 1.780 IB17A;
0.247 -0.417 1.020 KA73A;
0.295 -0.417 1.160 KB32A;
1.770 1.537 1.280 LA75A;
1.896 1.537 1.650 LB65A;
0.318 -0.204 0.770 NA120A;
0.069 -0.204 0.040 NA64A;
0.475 -0.204 1.230 NA71A;
0.428 -0.417 1.550 QB25A;
0.410 -0.417 1.500 QB61A;
0.297 -0.204 0.710 SB18A;
1.234 0.742 1.420 VB21A;

form standalone:
cat sequence_align.out
Interface residues:
chain_ID_A: A
residue_A:
chain_ID_B: B
residue_B:
Alignments:
Name chA chB ifResNum alnResNum linkscore ifSeqA ifSeqB

cat structure_align.out
Interface residues:

chain_ID_A: A

residue_A: 33 34 35 36 62 63 64 67 69 70 71 72 73 75 77 78 79 80 81 113 115 117 119 120 121 122 123 124

chain_ID_B: B

residue_B: 10 12 16 17 18 21 22 24 25 28 29 32 33 58 61 62 64 65 67
query A B 47 47 1.000 SAMDMTNIGANFKLDKGEPKEHSNDGSQ NDDISVLMQANKYQQSCLL


Alignments:
PDB chA chB ifResNum alnResNum score ifSeqA ifSeqB

----------------------------------------------------------------------------------------------------------------
the sequence_align.out file from standalone version seems not to show any sequences. Is this normal?
I am wondering If the problem with standalone version is the thing that you mentioned? If so I would appreciated it if you please tell me how can I fix it?
xiaoqiah
Posts: 6
Joined: Tue May 04, 2021 5:05 pm

Re: SSIPe

Post by xiaoqiah »

Hello, this is not normal.
The SSIPe will do the sequence alginment to generate the sequence_align.out file as well and the structure_align.out and sequence_align.out are used to generate the SSIP score. Because you did not obtain the sequence alignment, I guess there may be something wrong with the sequence alignment step. Can you make sure that you have constructed the STRING database as required? If you do not have the library, it is unlikely that you will get the alignment file.
Also I wonder if you have this kind of error for your 5M2O example only or any other examples? If the latter, then check the STRING database. If the former, the issues maybe somewhere else. You can reach me at my email xiaoqiah@umich.edu as I do not access this website very often.
Post Reply