Trouble demultiplexing dual barcoded, mixed orientation reads

[Michaelis_Vasiliadis]

Cutadapt seems to be discarding too many reads (~60%). We suspect this has to do with the mixed orientation flag. We have also spot checked a few reads from the demultiplexed file and there are some reads in each orientation. We have paired end sequencing with the same barcodes on each end. Removing the mixed orientation flag did not change the number of outputted reads. Any tips to improve this?

See example output from demux-paired command:
CGGCTYAATTYGAYTCAACRC (forward primer)
CGGCTTAATTTGACTCAACGCGGGGAAGCTCACCCGGCCCGGACACGGAAAGGATTGACAGATTGATAGCTCTTTCTCGATTCTGTGGGTGGTGGTGCATGGCCGTTCTTAGTTGGTGGAGTGATTTGTCTGGTTGATTCCGATAACGAACGAGACTCCGGCTTGTTAACTAGTTTCGCGGCCCCGTGTGGTCGGCGACTAACTTCTTAGAGGGACAAGTGGCGTTCAGCCACACGAGAAGGAGC
+

GGGCATCACRGACCTG (Reverse complement of reverse primer)
GGGCATCACAGACCTGTTATTGCTCCTTCTCGTGTGGCTGAACGCCACTTGTCCCTCTAAGAAGTTAGTCGCCGACCACACGGGGCCGCGAAACTAGTTAACAAGCCGGAGTCTCGTTCGTTATCGGAGTCAACCAGACAAATCACTCCACCAACTAAGAACGGCCATGCACCACCACCCACAGAATCGAGAAAGAGCTATCAATCTGTCAATCCTTTCCGTGTCCGGGCCGGGTGAGCTTCCCC

[lizgehret]

Hi @Michaelis_Vasiliadis,

Welcome to the forum!

Apologies for the delay in response on this - I am checking in with a few of my colleagues and we will circle back shortly. Thanks!

[lizgehret]

Hi @Michaelis_Vasiliadis,

Thanks for your patience! It looks like your primers are at the very start of both R1 and R2. If this is the case, you don't need to use cutadapt - you can pair (using vsearch, dada, etc) and then use --p-trim-left when running DADA2.

Try that and let us know if you are successful!