read splitting concatameric read Nanopore

[smartise]

Hello,

I recently did a nanopore pore run of metabarcoding read of COI amplicon. However due to measurment issue, too much DNA was input in the library and result in a large amount of concatameric reads.

I tried following multiple similar thread problems but fail to find a solution for my data. Basically my data would be as follow

BC1-COI-BC1-BC2-COI-BC2

and which to demultiplex it resulting in

BC1-COI-BC1 / BC2-COI-BC2

and then classified those read in different fasta file corresponding to the different barcode (barcode01.fasta, etc)

Is there any solution for me to do this ?

Thanks !

[marcelm]

Hi, this is currently not possible. One workaround would be to run Cutadapt twice, once to remove everything before the first BC2 sequence and once to remove everything after the second BC1 sequence. Then you could join the two output files and demultiplex the joined file by barcode.

This will only work for those reads that consist of two concatenated sequences. If your question is still relevant and you want to go that way, please let me know and I will help a bit more.

[smartise]

Hi, so in the end i managed to do it in a more simple way but without using cutadapt but minimap2. I build every possibilities of junction possible by combining every reversecomplement-forward. So it would be revcomp(BC1)-BC1, revcomp(BC1)-BC2 etc... for 96 barcode it resulted in a total of 9216 possibilities. i then mapped those possibilities using minimap2 and calculated the middle of each of those allignment. I then used a home made python script that did the split at this middle allignment and it worked quit efficiently. I didn't saved all my read but close to 90% i would say.

I definitly think that it could be added quit easily that way as it only took a couple line of code and worked pretty well