Structural perturbation of alpha-crystallin and its chaperone-like activity.

alpha-Crystallin is a multimeric lenticular protein that has recently been shown to be expressed in several non-lenticular tissues as well. It is shown to prevent aggregation of non-native proteins as a molecular chaperone. By using a non-thermal aggregation model, we could show that this process is temperature-dependent.

Sequence data analysis reveals a relationship between LSR2, the recombinant fusion protein mimicing M. Leprae and VIF of bovine immunodeficiency virus (BIV).

In the course of computer simulation study looking for active sites for the interaction between MHC II and T7--a 12 residue long peptide of LSR2--a recombinant fusion protein mimicing the native bacillus M.Leprae--an interesting relationship between the antigenicity of LSR2 and VIF of BIF has come to light. Computer analysis study has revealed this stretch of residue from 36 to 48 of LSR2 is highly antigenic.

Modified helix-loop-helix motifs of calmodulin--The influence of the exchange of helical regions on calcium-binding affinity.

The four calcium-binding sites, called the helix-loop-helix, or the EF-hand motifs, of calmodulin differ in their ion-binding affinities; this has been thought to arise due to the variations in the sequences of the loop regions where the ion binds. We focus attention here on the role of the flanking helical regions on the calcium-binding affinities. Peptides were synthesized in a manner that simulates the E and F helical flanks of site 4 (the strongest calcium-binding site of the calmodulin) to sandwich the loop sequences of sites 1, 2, 3 and 4 so as to produce peptides named 414, 424, 434 and 444, as well as using the helical flanks of site 1 (the weakest site) to produce peptides 111, 121, 131 and 141.

Recognition of Telugu characters using neural networks.

The aim of the present work is to recognize printed and handwritten Telugu characters using artificial neural networks (ANNs). Earlier work on recognition of Telugu characters has been done using conventional pattern recognition techniques. We make an initial attempt here of using neural networks for recognition with the aim of improving upon earlier methods which do not perform effectively in the presence of noise and distortion in the characters.