How to design an antibody with low background cross-reactions?

Chang Zhu czhu at
Tue Jun 18 01:28:38 EST 2002


The second question is definite more interesting. Unfortunately my
tool can not do that yet. What the software does is: giving a protein
sequence, find the most likely subregion that can be used to raise an
good antibody with few cross-reactions. The algorithm does not predict
structure, so will be far off to answer your questions. But you are
right in asking these questions: if one has an accurate
structure-based algorithm, we will be one step closer to predict
protein-protein interactions.


"Emir Khatipov" <khatipovNO at> wrote in message news:<mStP8.45$V4.8942 at>...
> Chang,
> Can your tool (or its extension) predict the sequence of the antibody
> binding site or, what is more interesting to me, any peptide sequence that
> would selectively bind a certain region of the target protein?
> Emir
> "Chang Zhu" <czhu at> wrote in message
> news:5aa28508.0206171011.6c2f7050 at
> > A common problem in raising an antibody is the high background
> > cross-reactions to non-target proteins. One of the reasons is that
> > proteins share similar sequences. Certain sequences are
> > over-represented. For an example, searching a non-redundant mouse
> > protein database I found that the quadri-peptide SGSG is present in
> > the mouse proteome for 5197 times in 3170 proteins, and the
> > quadri-peptide ACWM is present for only 2 times in 2 proteins. If the
> > sequence you used for raising an antibody happens to contain an
> > over-represented sequence, it is likely a fraction of the antibody
> > will reacts with the over-represented sequences, therefore
> > cross-reacts with a large number of non-target proteins.
> >
> > Chang Bioscience has introduced Abie Pro 3.0 to reduce this source of
> > cross-reactions. Abie Pro searches a proteome for over-represented
> > sequences and gives a user an option to select only peptides with few
> > over-represented sequences for antibody production. The online version
> > is available at
> >
> >
> >
> > You are invited to try it. Critics, comments, and suggestions will be
> > especially appreciated.
> >
> > Thanks.
> >
> > Chang

More information about the Methods mailing list