No announcement yet.

Problem with indexes/demultiplexing

  • Filter
  • Time
  • Show
Clear All
new posts

  • Problem with indexes/demultiplexing

    Dear all,
    I have a question about how illumina NextSeq 550 does the demultiplexing. I have looked into the demultiplexing summary files and I can see that the most popular indexes are the ones that I used (and introduced in the sample sheet, with the correct sample names), but then I can see a bunch of other indexes, that do not match the other indexes available in my RNA library preparation kit. They are less frequent, but still have a 10% less of hits as the "original" indexes. So for example, one of the used indexes has 2698225, but one of the "random" indexes have 227303. Is this normal? More importantly, are these "random" indexes considered as real indexes and thus included as an independent read and then classified as to the index that has the highest similarity?
    In my fastq.gz files I have my 9 samples and the unfiltered ones. I am just afraid that there was some mistake during the demultiplexing and so the reads are mixed in the samples.

    Thanks a lot.

  • #2
    Reads are always lost to "random" indexes in multiplexed pools. Some are sequencing/phasing errors. There is not much one can do about these. As long as you recovered expected indexes in majority of data that is fine.

    You can allow up to 2 errors (if your indexes allow for it) to recover additional data by accounting for sequencing errors. If this has not been done you can at least try 1 error and re-demultiplex the data to recover some additional reads.