BLAST Output :gene prediction

[parimaladevi]

I'm working on BLAST with Neisseria meningitidis with around 2500 genes.
For my sample, I am getting double the number of genes predicted than the ab-initio gene prediction tools, and I'm working from command line.

How do I use culling limit?
Is there any option to avoid repetitive and false positive results?


The commands used:

Making the database:
makeblastdb -in finaldb.fa -dbtype 'nucl' -out final

Command:
blastn -db final -query input.fa -outfmt 6 -out output

[mastal]

What are you using as your database, how many sequences in the database, and what are you using as the query?

If you are running blast from the command line, type 'blastn -help', and you will get the online help page with the list of parameters for blast.

You can get fewer hits by setting a lower e-value, or a higher percent identity.

[bastianwur]

What's the problem with getting twice the amount of genes with blasting?
I mean BLAST will show you still a good hit even if start and stop codon are not in it (means there's no gene, but a good blast hit).