You are here

antibody modeling fail

2 posts / 0 new
Last post
antibody modeling fail
#1

Hi: 

https://rosie.graylab.jhu.edu/antibody/viewjob/72943

anyone please have a look this error

Thanks a lot

Jessica

Post Situation: 
Tue, 2019-07-02 01:57
jessye

Your issue is that the antibody modeling protocol requires certain sequence motifs in order to identify the CDRs. These are based on conserved residues typically found in the framework regions of the antibody.

Specifically, the server is having issues identifying the LCDR1 region (hence the error message: "UnboundLocalError: local variable 'L1_end' referenced before assignment" - this is a somewhat obtuse way of saying that the protocol can't find (i.e. assign) the end of the L1 CDR.).

The pattern that it's looking for to locate the LCDR1 is a conserved cysteine, followed 2-18 residues later by a consereved tryptophan in a particular three amino acid sequence (WYL|WLQ|WFQ|WYQ|WYH|WVQ|WVR|WWQ|WVK|WYR|WLL|WFL|WVF|WIQ|WYR|WNQ|WHL|WHQ|WYM|WYY).

Your light chain doesn't have this motif -- in fact, it's missing the conserved cysteines altogether. Unfortunately, you wouldn't be able to use the ROSIE antibody modeling protocol as-is with this sequence. The best you're likely to be able to do is to attempt to model a closely related sequence which does have the conserved disulfide cysteines. (My guess is that this is an antibody which has been manually engineered to remove the disulfides. It should be possible to take a look at where this sequence came from, determine which residues were mutated away from the conserved cysteines, and then just put the cysteines back for modeling purposes.)

Wed, 2019-07-10 09:09
rmoretti