I'm working on a genome project and using an in silico k-mer analysis to estimate the size of our genome based on the available Illumina reads. The k-mer based estimate is consistent across a wide range of k values, but is substantially lower than the previous estimate based on flow cytometry (described in Experimental procedures of this paper). How accurate are estimations of genome size based on flow cytometry?

  • $\begingroup$ Could you please edit your question and provide information on how exactly the genome size was determined by flow cytometry? $\endgroup$
    – Eekhoorn
    Commented Jul 7, 2014 at 15:02
  • 1
    $\begingroup$ If I understand the paper correctly, they are using fluorescence of nuclear DNA to estimate the genome size, presumably using a standard curve control, which is very clever and I guess if they do not have any saturation problems in signal detection and if their assumption about fluorescence to DNA size linear relationship holds true (presumably having done a standard curve), and if dense chromosomal areas don't make some kind of odd fluorescent effects then I don't see a major flaw in their work but very interesting and at the same time very simple and potentially high-throughput! $\endgroup$ Commented Jul 7, 2014 at 18:56
  • 2
    $\begingroup$ I don't think your results conflict. k-mer analysis is a measure of genome length, and flow cytometry is a raw measurement of DNA quantity. If I understand k-mer analysis correctly, a diploid or polyploid organism will give the genome length under k-mer analysis, and the genome length * the ploidy under flow cytometry. Are your results close to an integer multiple of their results? $\endgroup$
    – Resonating
    Commented Jul 7, 2014 at 19:32
  • $\begingroup$ @JeremyKemball No, it's not near an integer multiple, so it's not a ploidy issue. The reasoning behind the k-mer analysis is to identify a sequence that occurs once in the genome and count how many times it occurs as a proxy for coverage. Since that unique sequence is unknown, we instead examine k-mer distributions and look for the mode in those distributions. $\endgroup$ Commented Jul 8, 2014 at 14:45
  • $\begingroup$ Perhaps highly repetitive regions in the genome? Those should stand out in your k-mer analysis, though. If the reads have been filtered to remove hard-to-assemble repetitive regions, that might explain the discrepancy. $\endgroup$
    – Resonating
    Commented Jul 8, 2014 at 15:17


You must log in to answer this question.

Browse other questions tagged .