[Zbrafish] Re: Sequence homology for protein expression

[Wilson Clements]

Re: Sequence homology for protein expression

Dear E.A.,

I would say that to be careful, you should consider different amino  
acid substitutions in different clones to be a problem.  For the  
following reasons:

1) The fact that you are not getting the same errors clone-to-clone  
strongly suggests that you are getting PCR errors.  High fidelity  
enzymes are not perfect.  Are you running high numbers of  
amplification cycles (>30-32)?  The less cycles, the less errors you  
should get, although you can still be unlucky and get errors at low  
numbers of cycles.  You could try:  using a variety of enzymes, I now  
prefer Phusion (from Finnzymes) and iProof (from Bio-Rad), they seem  
to give me the lowest number of errors.  Making all your cDNA from a  
single clutch from a pairwise cross.  Then you will know that you  
should get very few polymorphisms clone-to-clone.

2) Many amino acid substitutions don't matter, but without having the  
"wild-type" to compare functionality against, it will be hard to be  
sure yours doesn't.  Especially a Threonine to Alanine substitution,  
since Thr is frequently a phosphorylated amino acid for protein  
regulation.  If I start getting "polymorphisms" different from the  
expected amino acid sequence, I sequence quite a few more clones to  
make sure they all have the same substitutions.  I don't worry so  
much about silent base substitutions (in the third position of the  
codon leading to no change in amino acid).

In general, I don't think pairwise BLAST is a very good way to assess  
whether your clones are good.  The BLAST algorithm tends to cut off  
ends or regions that are "not conserved" so you don't get a base-by- 
base comparison.  We use the Sequencher software here.  I'm not sure  
if you have access to it, but many universities and companies have  
purchased networked licenses that you can use on a client/server  
basis using downloaded software.  I think it's the best thing going.   
You can make a "virtual" clone of your expected sequence (in the  
vector) and then do an alignment (make a contig) against the actual  
sequencing returned.  Then you can look base-by-base at which bases  
match and which do not.  In addition, you can consider the quality of  
the sequencing data.  If your "polymorphisms" are towards the end, or  
very beginning, of a sequencing run, the sequence quality may be  
pretty bad.  If so, you're clone could be correct and you're getting  
fooled by bad sequence.  You can't assess this parameter doing a  
pairwise BLAST.  If you don't have Sequencher, I would still  
recommend trying to find some sort of sequence analysis software,  
there's quite a bit out there for free.

Good luck.

Wilson
------------------------------------------------------------------------ 
--------------------
Wilson Clements, Ph.D.

wclements from ucsd.edu

Dept. of Biology
Section of Cell and Developmental Biology
University of California at San Diego
9500 Gilman Dr.
Natural Sciences Building 6105
La Jolla, CA 92093-0380

TEL    (858) 534-6955
LAB    (858) 822-4658
FAX    (858) 822-5740


Begin forwarded message:

> Hi Joanna,
>
> Thanks for the answer. The strange thing is that I do not get the same
> ‘errors’ when sequencing multiple clones. I can only think that this
> is due to the fact that the cDNA was made from multiple larvae
> ? Or
> that although I used HF the enzyme is causing errors. In any case,
> when I use blastx for protein sequence (part of the protein – 900bp) I
> get something like this:
>  Score =  446 bits (1147),  Expect = 1e-123
>  Identities = 220/221 (99%), Positives = 220/221 (99%), Gaps = 0/221
> (0%)
>  Frame = +2
>
> (One alanine substituting a threonine)
> Is such a substitution a problem?
> Thanks very much
> E.A.
>

-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://www.bio.net/bionet/mm/zbrafish/attachments/20080423/c2159b98/attachment.html