Filtering raw sequence - mismatch with Ns

[jmugford]

I have Illumina reads which I have barcoded prior to adding the Illumina adaptors. I now need to sort the raw reads into separate files before mapping. I'm using perl and have everything running smoothly if the barcode matches the raw sequence perfectly. My problem comes in when there is a mismatch or N called within the barcode sequence. Does anyone have an idea of how to efficiently code for this?

Many thanks.

[arvid]

You could check out the idea used in Novocraft's novobarcode (http://novocraft.com/main/page.php?s...code_userguide), where they use the alignment quality to allow for mismatches. Usually barcodes are choosen carefully to allow for clear distinction even though you have one or two mismatches... If you don't want to code it yourself you can use their tool, it is free for academics.

[twaddlac]

I've done it by using a mapper's output (nucmer/bowtie) to get the coordinates and percent identity and then filter/trim/sort based on those. It makes for easy scripting since they're tab delimited.

[adamdeluca]

I have used a tool from expressionAnalysis called fastq-multx. It worked pretty well for me.

http://code.google.com/p/ea-utils/