Error when using deML with dual indexed paired end sequencs

Hey, I've been trying to get deML to demultiplex some dual indexed sequence data. However, I keep receiving an error stating that their are numerous conflicts in my index filed

I've been running this

src/deML -i Beetle_Index.txt -f 2226_S1_L004_R1_001.fastq -r 2226_S1_L004_R4_001.fastq -if1 2226_S1_L004_R2_001.fastq -if2 2226_S1_L004_R3_001.fastq -o Data/demultiplexed.

and keep getting this error (I've just listed the top few lines):

Conflicts for index1:

TCCGGAGA from CarK_2016 causes a conflict with MinK_2016 FulK_R2_2016 
CGCTCATT from PitchB_2016 causes a conflict with WobB_2016 CoyB_R2_2016 
GAGATTCC from Win_B_2016 causes a conflict with CarK_2016 PigB_R2_2016 
ATTCAGAA from SweetB_2016 causes a conflict with MusB_2016 OakB_R2_2016 
GAATTCGT from MinSB_2016 causes a conflict with CarB_2016

Which I can understand since my Index file looks like this:

#Index1          Index2            Name
TCCGGAGA    AGGCTATA    CarK_2016
CGCTCATT    AGGCTATA    PitchB_2016
GAGATTCC    AGGCTATA    Win_B_2016
ATTCAGAA    AGGCTATA    SweetB_2016
GAATTCGT    AGGCTATA    MinSB_2016
CTGAAGCT    AGGCTATA    WobB_2016
TAATGCGC    AGGCTATA    PigB_2016
CGGCTATG    AGGCTATA    WilK_2016
TCCGCGAA    AGGCTATA    WinB_2016
TCTCGCGC    AGGCTATA    CedB_2016
AGCGATAG    AGGCTATA    FulB_2016
ATTACTCG    GCCTCTAT    HeavK_2016
TCCGGAGA    GCCTCTAT    MinK_2016
CGCTCATT    GCCTCTAT    WobB_2016
GAGATTCC    GCCTCTAT    CarK_2016
ATTCAGAA    GCCTCTAT    MusB_2016
GAATTCGT    GCCTCTAT    CarB_2016
CTGAAGCT    GCCTCTAT    AprC_2016
TAATGCGC    GCCTCTAT    WinB_R2_2016

Obviously there are repeats in index2 but the point is for their to be a unique combination when paired with index 1, so that shouldn't matter. Is there some Option I'm not invoking that will make deML pick up on this fact? Or have I formatted the index file incorrectly?

sequencing next-gen demultiplexing deml

• 2,283 views • link

updated 7.4 years ago by Gabriel R.

• written 7.8 years ago by drunva •

Hey @drunva

Did you manage to figure this out? I am in a similar scenario where there are lots of conflicts but no errors per say.

Thanks!

• 0 views • link

written 7.4 years ago by aguitarfreak •

1 answer

Hi both, those are just warnings. deML (http://grenaud.github.io/deML/) is robust to poorly designed indices lists or sequencing runs with very poor quality. These warnings just inform you that an index was used. I have since added this to the README, let me know if this is sufficiently clear and verbose:

I get "ACGTACGT from sample1 causes a conflict with sample2" why?

These are just warnings, not necessarily errors if you have double indexing. If you have single indices, this means the algorithm cannot distinguish between "sample1" and "sample2" and all reads assigned to those will fail internal QC.

For double-indexing, it merely indicates that in your index list, you have some repeated indices for either the first or second index. Of course, having unique pairs is sufficient to demultiplex. However, having repetitive indices for one index defeats the purpose of using double-indexing as you rely on a single index to demultiplex. Again, this might be unavoidable if your sample list is greater than your list of unique indices.

• 0 views • link

written 7.4 years ago by Gabriel R.