Header Leaderboard Ad

Collapse

Dots in the reads of SoliD sequencing

Collapse

Announcement

Collapse
No announcement yet.
X
 
  • Filter
  • Time
  • Show
Clear All
new posts

  • Dots in the reads of SoliD sequencing

    Hi everyone!

    I am a bioinformatics student and I have just started dealing with SOLiD reads.

    The sequenced reads I have are full of dots ("."). This is an example:

    >1_459_10_F3
    T1.0212.2323.2320321323213320323212213...20...20...
    >1_480_41_F3
    T1.1213.2223.1320112233101130002201222...11...03...
    >1_482_41_F3
    T1.0111.1023.2320110113122032121122101...21...13...

    I have read that those dots are bases that could not be identified and, when translated, would be "N"s.

    My question is how to deal with this reads? Discard the entire read? Trim the read? (I think trimming would generate very short reads that would not be good for alignment.)

    Does anyone knows how the aligners (bowtie, novoalign, maq, soap, etc) deal with it?

    Thanks in advance!

  • #2
    Check how many there are. I have seen dots in SOLiD reads mostly towards the start of the file.

    If they all, or a significant proportion of them, have dots then I would doubt the sequence quality. Perhaps talk to your provider about this.

    Re trimming, it would be fine to trim the last 10 bp off or so, but more, as you say, could affect the alignments.

    Comment


    • #3
      Thanks for your answer colindaven!!

      I have checked the file and all the reads have at least two dots in them, mostly in the beginning. The qualities for those positions are all negative values, which apparently is the value for unknown bases.

      I have talked with the group that did the sequencing and they said they had trouble analyzing this data.

      I am also wondering how much I can trust in the bases after the dots, considering solid's coding and double checking....

      Comment


      • #4
        Why not try aligning to your reference sequence (I hope you have a good one!).

        Then try removing the first say 5-10-15bp (depending on where the dots are) and realigning.

        And, if I were you, I would talk to those responsible about the data quality.

        SHRIMP is a decent open source aligner for mapping if you don't want to go for NovoalignCS or Lifescope.

        Comment


        • #5
          Thanks again for you answer!

          That's a good idea! I'll try comparing the alignments.

          Actually, I'm using NovoalignCS for my tests. I know that it is a good aligner, but I don't know how it deals with the dots. I've read the manual but I could not find it? Do you know where I can get this information?

          Comment


          • #6
            Hey Nnogueira,

            How did you end up treating these reads? did you trim them at the ends, discard the ones with dots?

            Any input would be greatly appreciated, as I am now in your situation!

            Best,
            Carmen

            Comment


            • #7
              Hi carmeyeii,

              we are developing a software to deal with this kind of error. Probably, it will be available within a month.

              I will let you know if have any news.

              Comment

              Working...
              X