Prediction of protein secondary structure by using local alignments.

Method is based on comparison of charcteristics, calculated for positions of processing sequence, such as aminoacid exposure to water, submergence of aminoacid residue into molecule body etc, with the same characteristics, obtained from analysis of PDB-files in database.

FASTA formatted sequence or specially prepared alignment (see example) can be used as an input. The number of aligned sequences must be less than 250 !!!

Input sequence for this program should be in fasta format with 80 or less sequence letters per line.

Accuracy

Overall 3-state (a, b, c) prediction gives about 75% correctly predicted residues. THIS ACCURACY IS REACHED WITHOUT USING MULTIPLE ALIGNMENT INPUT when it is higher SEE ALSO "SSP" and "NNSSP" programs.

Output results with probability of prediction:

>T0197                                                                
Length= 179  4849   27.1
                    10        20        30        40        50
 PredSS     aaa        aaaaaaaaaaaaaaaa    aaaaa     bbbb    b
 AA seq     AHHHHHHGSEVEIKFKIKLEDFLHTLNTFNPEFVRYEEQEDVYFEVPRPK
 ProbA      86621110023345434467789876433134555333532111110010
 ProbB      11233210135543452211100111211003333211123667732148
                    60        70        80        90       100
 PredSS     bbbb     aaaaaaaaaaaaaaaaaaaaaaaaa   aaaaaaaaaaa  
 AA seq     LLRIRGVHNLKKYYLTFKEILDENNEEFYEVEFEIGDFEKAVEVFKRLGF
 ProbA      00112332257899979999888878887796643227999999999422
 ProbB      88753111111100010000000000001101222010000000000001
                   110       120       130       140       150
 PredSS     bbbbbb   bbbbbb  bbbbb bb        bbbb   aaaaaaaaaa
 AA seq     KIQATIKKKRWVYKLNGVTLEVNRVEGIGDFVDIEVISDSPEEAKEKIWE
 ProbA      22112222322112111100001111111111000111116788999999
 ProbB      55665521256776521667984751000034499763110000000000
                   160       170
 PredSS     aaaaa        aaaaaaaaaaa     
 AA seq     VAKMLGLKEEDVEPRLYLELINELSGRSS
 ProbA      99996233323325799999998632211
 ProbB      00001111110101100000000100122

• 1 line - sequence name
• 2 line - length of protein and maximal scores of similarities with DB sequences
• PredSS - secondary structure predicted
• AA seq - query sequence
• ProbA - a row probability of the position to be in alpha helix
• ProbB - a row probability of the position to be in beta strand

An example of alignment (numbers of positions are not necessary):

     ACTINOXANTHIN                                                         
         5  107
             10        20        30        40        50        60 
     APAFSVSPASGASDGQSVSVSVAAAGETYYIAQaAPVGGQDAaNPATATSFTTDASGAAS 
     APAFSVSPASGLSDGQSVSVSGAAAGETYYIAQCAPVGGQDACNPATATSFTTDASGAAS
     APTATVTPSSGLSDGTVVKVAGAgaGTAYDVGQCAWVdgVLACNPADFSSVTADANGSAS
     APGVTVTPATGLSNGQTVTVSATgpGTVYHVGQCAVvpGVIGCDATTSTDVTADAAGKIT
     ATPKSSSGGAGASTGSGTSSAAVTSgaASSAQQSGLQGATGAGGGSSSTPGTQPGSGAGG
             70        80        90       100      
     FSFTVRKSYAGQTPSGTPVGSVDbATDAbNLGAGNSGLNLGHVALTF
     FSFV-RKSYAGZTPSGTPVGSVDCATDACNLGAGNSGLNLGHVALTF
     TSLTVRRSFEGFLFDGTRWGTVDCTTAACQVGLSDAAGNGpgVAISF
     AQLKVHSSFQAVvaNGTPWGTVNCKVVSCSAGLGSDSGEGAAQAITF
     AIAARPVSAMGGtpPHTVPGSTNTTTTAMAGGVGGPgaNPNAAALM-

• 1 line - sequence name
• 2 line - number of aligned sequences and length of protein
• 3 and subsequent lines - aligned sequences in format 60a1

You can use small letters for amino acids, if you want.
Alignment MUST be without deletions in the 1-st (query) sequence!!!

Reference:

Salamov A.A., Solovyev V.V.
Protein secondary sturcture prediction using local alignments.
J.Mol.Biol.1977, 268,1, 31-36.

Salamov A.A., Solovyev V.V.
Prediction of protein secondary sturcture by combining nearest-neighbor algorithms and multiply sequence alignments.
J.Mol.Biol.1995,247,1,11-15.