Hi everyone! I'm now looking for some off-target sites of a 36bp-DNA sequence. But I can't get positive results. I use N for replacing some of the letters in the sequence to do blast on NCBI. (e.g. ATTTGCC, I use NTTTGCC, ANTTGCC,..., ANNTGCC etc.). This method is very laborous. Are there any convenient way or some web sites in which I can do the prediction? Thank you.
Best regards
Stephen
0
3 replies to this topic
#2
pcrman
-
- Global Moderators
-
- 1,031 posts
Epigenetist
46
Excellent
Excellent
Posted 04 June 2010 - 11:00 PM
One way is to set a big Expect value (10) in NCBI BLAST to get more hits.
It seems that Enseml BLAST program gives you more choices. For example, you can use a E value of 100000 which gives you significant more hits than a E value of 10 in NCBI.
It seems that Enseml BLAST program gives you more choices. For example, you can use a E value of 100000 which gives you significant more hits than a E value of 10 in NCBI.
#3
Jin
-
- Members
-
- 4 posts
member
0
Neutral
Neutral
Posted 09 June 2010 - 04:00 AM
pcrman, on Jun 5 2010, 03:00 PM, said:
One way is to set a big Expect value (10) in NCBI BLAST to get more hits.
It seems that Enseml BLAST program gives you more choices. For example, you can use a E value of 100000 which gives you significant more hits than a E value of 10 in NCBI.
It seems that Enseml BLAST program gives you more choices. For example, you can use a E value of 100000 which gives you significant more hits than a E value of 10 in NCBI.
Thank you pcrman. I'll have a try.
#4
HomeBrew
-
- Global Moderators
-
- 927 posts
Veteran
14
Good
Good
Posted 09 June 2010 - 04:33 AM
I'm not sure I understand what you're doing...
If you're trying to find all occurrences of a query pattern (say ATTTGCC) in a larger target sequence, and allow such a match to be counted even if it's off by one or more characters, you should try the fuzznuc program included in the EMBOSS suite, setting the -pmismatch switch to however many mismatches you want to allow. You should also set the -complement switch to true if you want to search both strands of the target sequence for the query pattern.
If you're trying to find all occurrences of a query pattern (say ATTTGCC) in a larger target sequence, and allow such a match to be counted even if it's off by one or more characters, you should try the fuzznuc program included in the EMBOSS suite, setting the -pmismatch switch to however many mismatches you want to allow. You should also set the -complement switch to true if you want to search both strands of the target sequence for the query pattern.
