Your question may be too broad to answer adequately. Finding unique patterns in a gene sequence doesn't necessarily reveal much if you're trying to compare a potentially orthologous sequence between humans and some non-human species. You'd need to know something else about the sequence - for example, you can explore metabolic pathways across species by observing which genes are present or absent. You could also go hunting for mobile genetic elements in a similar way. The options are essentially limitless.

Hello Harry, Thanks for your reply. In my work actually I need to find out some unique stretches of a bacterial gene (window length can be 20-30) in which those should not present in human genome. Those unique sequences will be used in further studies.

initially I have done BLAST for the gene against human genome but i did not find any matches, later i tried with 20-30 nucleotide length by changing the parameters (E-value, word size, gap penalty) this time I am able to see matches for 15-25 nucleotide length in human genome with high E value and gaps. But I want to know is there any other method or algorithm to directly align two sequences to get unique patterns.

You need to be more specific in what you are looking for. If you are looking for a particular pattern, the programming language perl is great for parsing data into usable sets for analysis. Or try using some basic unix commands to figure out what youre looking for. If you could give us a little more information other than pattern, we can provide better answers. Bacterial genes are pretty well characterized, and most if not all of the active domains are understood. If you are looking for protein binding domains in promoter, intron or intergenic regions, then say so. There are plenty of programs that can help you with motif identification. If you cannot be more specific, I cannot imagine you'll get more specific answers. Sorry for the double negative there.

Hello Alexandre, I am not looking for a particular patterns. I am looking for a unique stretch of sequence of a bacterial gene say (AGGTTC of 20-30 nucleotide length) that should not present in human genome say (AGTGTGCGAACTCC), where the unique stretch of bacterial gene cannot match any where in human genome(it may align with some gaps here and ther but not entire stretch).

Any bacterial gene or one in particular? There is a very easy way to do this if you can write in perl. Your basically trying to do a blast K-mer analysis but in reverse. So if you don't know the sequence, your best bet is to take all the various 20mers from the bacterial transcriptome, then ask perl to parse those sequences by asking if they match the human genome or not. If you can learn perl, this is very simple. There is no program currently available to my knowledge that will do this for you. I hope this has helped.

So are you trying to find something like the results of this 2011 Anderson and Seifert paper (http://mbio.asm.org/content/2/1/e00005-11.full) ?