| 主权项 |
1. A method for the production of a polypeptide of interest by a eukaryotic host cell, which method comprises modifying a value of a set of relevant protein features in an amino acid backbone of the polypeptide to fall within an optimal range or to become closer to an optimal value for one or more protein features in the eukaryotic host, wherein said protein features are properties that can be calculated from the protein amino acid sequence and DNA sequence, the method comprising the steps of:
(i) providing a sequence for the polypeptide of interest, (ii) determining an optimal range and an optimal value for one or more protein features in the eukaryotic host, (iii) determining a set of relevant protein features in the eukaryotic host, which features will improve secretion of the polypeptide by the eukaryotic host if one or more of said relevant features is modified in the amino acid backbone of the polypeptide of interest, and (iv) modifying the sequence for the polypeptide of interest such that the value of the relevant protein features falls within the optimal range or closer to the optimal value as determined in (i), and (v) producing the modified polypeptide by the eukaryotic host cell, wherein said method improves the secretion of said polypeptide of interest, wherein the specificity and/or specific activity of the polypeptide after said modifying substantially remains the same as before improvement of the secretion, wherein (i) and (ii) may be performed in any order, wherein the relevant set of protein features is determined by:
a. collecting or creating a dataset S, which contains secretion levels of a suitable amount of proteins in a certain eukaryotic host and amino acid and DNA sequences thereof;b. calculating values of protein features (F) for all proteins in the dataset S;c. using a statistical classification method to select a subset of protein features (Fs) that gives the best performance of a statistical classifier to distinguish between secreted proteins (S+) and non-secreted proteins (S−) in the dataset S, according to a suitably defined classifier performance criterion; andwherein the optimal range or the optimal value of protein features for a eukaryotic host is determined by:d. collecting or creating a dataset S, which comprises secretion levels of a suitable amount of proteins in a certain eukaryotic host and amino acid and DNA sequences thereof;e. calculating protein features (F) for all proteins in the dataset S;f. determining an optimal value (F opt) for each feature for the eukaryotic host by fitting a probability distribution for each protein feature calculated from secreted proteins (S+) such that the distribution of the feature values is well described by the chosen probability distribution;g. determining an optimal range of each feature for the eukaryotic host;wherein the relevant set of protein features is selected from the group consisting of: the number of amino acids, molecular weight, isoelectric point, net charge at a specific pH, GRAVY score, aliphatic index, instability index, compositional features, atomic composition of C, H, N, O, S atoms, amino acid frequency, dipeptide frequency, tripeptide frequency, acidic amino acid frequency, aliphatic amino acid frequency, aromatic amino acid frequency, basic amino acid frequency, local features, localization features, glycosylation pattern and/or charged amino acid frequency. |
