Kyte-Doolittle

Uversky and co-workers recently used a pair of sequence attributes, specifically the Kyte-Doolittle hydropathy scale and net charge, to... 

Figure 14. Hydrophobicity analysis of five P-type ATPases according to the Kyte-Doolittle method. A hydrophobicity value between -4.5 and +4.5 is assigned to each type of amino acid residue and mean values are successively calculated along the peptide sequence using a window of 18 residues. Segments corresponding to the transmembrane helices M1-M10 in the structural model in Figure 15 are shaded. Modified from Verma et al., 1988.

Volume = Volume enclosed by van der Waals radii Mass = molecular weight of nonionized amino acid minus that of water both adopted from Creighton (1993) HP scale = degree of hydrophobicity of amino acid side chains, based on Kyte Doolittle (1982) Surface Area = mean fraction buried, based on Rose et al. (1985) and Secondary structure propensity = the normalized frequencies for each conformation, adopted from Creighton (1993), is the fraction of residues of each amino acid that occurred in that conformation, divided by this fraction for all residues. 

Table 5.2 List of the cytochrome c fragments contained in the commercial digest sample with their sequence, mono-molecular weight, isoelectric point (pi) and hydrophobic character (log P, Kyte-Doolittle scale) calculated using Expasy software. ...

By using the cross-validation statistical procedure and Kyte-Doolittle hydropathy scale, the prediction results for TMH in the training data base of 63 membrane proteins common to us and to Rost et al. [9] and also to Jones et al. [33] were similar in accuracy by all three methods. When training data base is enlarged to 168 proteins, we maintain the 95% accuracy for predicted transmembrane helices and almost 80% (78.6%) of proteins are predicted with 100% correct transmembrane topology. When 168 proteins are divided in the above mentioned training set of 63 proteins and an independent test set of 105 proteins, all performance parameters for TMH prediction associated with a set of 105 proteins exhibited a decrease which was smaller in our case than for Rost et al. [9]. 

The SPLIT algorithm was optimized for predicting transmembrane a-helices by using the Kyte-Doolittle hydropathy scale to create profile of a-helix preferences. The digital version of prediction for transmembrane a-helices is designated as the TMH predictor. Predicted profile of P-strand preferences can be used to find sequence location of potential membrane-embedded or surface-attached P-strands. The score for potential membrane-attached P-strand... 

The prediction results with Kyte-Doolittle preference functions... 

Complete prediction results for the gef ecoli protein by using the Kyte-Doolittle hydropathy... 

Ten integral membrane proteins of well known structure (BESTP, Methods) have been tested first. Only the Kyte-Doolittle and our modification of the Kyte-Doolittle scale (MODKD, 83) were able to predict all od these ten membrane proteins with 100% correct transmembrane topology, i.e. all transmembrane helices were correctly predicted at their observed sequence locations and there were no overpredicted TMH (Table 7). Only the Chothia buried surface scale (CHOTH, 29) did not recognize one of ten membrane proteins as the membrane protein (the subunit H fi-om the photosynthetic reaction center from R. viridis). Nine long extramembrane helices in these 10 proteins were not predicted as TMH by any of 12 tested amino acid scales. That these sensitive tests of our predictor do not depend on the chosen training procedure was checked by using different training procedures. After... 

Figure 3 Score profiles for cxlbjarde (Figure 3A) and for cox3 parde (Figure 3B) of cytochrome oxidase from Paracoccus denitrificans [14] are obtained by substraction of turn preferences from a-helix preferences (full line). Digital predictions, as outcome of the best training procedure for the SPLIT algorithm with Kyte-Doolittle hydropathy scale (Methods), are shown as bold horizontal bars at the score level 0.5. Observed location of TMH segments are shown as bold horizontal bars at the score level 0.2.

The Kyte-Doolittle scale is used in each case See Methods for performance parameters. 

Figure 4 Score profiles for porin from Rhodobacter capsulatus are obtained by subtraction of turn preferences from helical preferences (full line) and as sum of P-sheet preferences and hydrophobic moment scores for assumed p-sheet conformation (dotted line). Kyte-Doolittle scale [17] is used to calculate preferences, while PRIFT scale [50] is used to calculate hydrophobic moments. Observed transmembrane strands are shown as bold horizontal bars at the score level 2.0.

Our algorithm can give partial answer to the question what attributes are optimal predictors for specific folding motifs. Kyte-Doolittle type hydropathy values and Chou-Fasman type conformational preferences are two obvious answers to the question what amino acid attributes are good predictors for majority of transmembrane helices. Indeed, three such scales MODKD, KYTDO and CPREF (Table 4), are on the very top of the list of the best amino acid scales (Table 5). Performance parameters that punish overprediction (A-j y and Qp) give advantage to hydropathy values. Modifications to the Kyte-Doolittle values in the MODKD... 

Well known Kyte-Doolittle scale [17] can be used throughout, except in the case when specific need exists to test other amino acid attributes. 

Two data bases of soluble proteins of known structure used to find false positive prediction results (Table I and Table II). Gaussian parameters needed for evaluation of preference functions based on the Kyte-Doolittle hydropathy scale [17] (Table III). Table with detailed prediction results for transmembrane helices in 168 integral membrane proteins (Table IV). Table with a detailed comparison of prediction results for 10 best known membrane proteins for our and three other algorithms (Table V). All these tables together with the FORTRAN 77 source code are available from the anonymous ftp server mia.os.camet.hr in the /pub/pssp directory. The anonymous login is ftp and the e-mail address is accepted as password. The list of files with short descriptions is contained in the 00index.txt file. 

The best known and most used procedure is the Kyte-Doolittle method which computes within the sliding window of specific width the hydrophobicity/am-phiphilicity of the segment. This represents a certain probability that the specific segment will or will not be present in the membrane. 

It can be seen from the AflX) profiles shown for several proteins in this report that the profiles in the present version of the PREF method are much clearer than in the earlier version. These profiles are far clearer than those obtained by the Kyte-Doolittle method on the example of cyoe ecoli. ... 

Figure 11.2. Results of a Kyte-Doolittle hydropathy determination using TGREASE. The input sequence was of the high affinity interleukin-8 receptor B from human. Default window lengths were used. The thick, horizontal bars across the bottom of the figure were added manually and represent the positions of the seven transmembrane regions of IL-8R-B, as given in the SWISS-PROT entry for this protein (P25025).

Analyses of these proteins with the TMPredict algorithm and comparison of these outputs to the Kyte-Doolittle hydrophobicity plots indicate that these proteins are... 

---