با استفاده از شبكه عصبي مصنوعي DNA تعيين ساختارهاي موثر در اتصال آنتي باديهاي به

Determination of effective structures of antibodies in binding to DNA using artificial neural networks

The accumulation of DNA-antibody immune complexes in kidneys is associated with SLE in human and lupus in mice. The level of serum of anti-DNA antibodies is also directly related to the disease severity. Determination of structures that confers DNA specificity in these antibodies can help in designing new therapeutic strategies. In this study, computer simulation and artificial neural network (ANN) has been utilized for the first time to determine effective structure of anti-DNA antibodies in binding to DNA. VH CDR3, FR3 and CDR2 were designated the most effective regions in antibodies that bind to DNA, using an educated neural network. We have also shown that the presence of Arg in position 104, 109 and 113 of CDR3, position 63 and 64 of CDR2 and position 90 of FR3. enhances the DNA specificity of antibodies. A part of our result confirms the previous laboratory works, and the other part introduces novel structures that may be important in DNA binding. This study shows the potency of logic networks in solving medical dilemmas.