View Single Post
Old 11-06-2015, 10:38 AM   #2
Brian Bushnell
Super Moderator
Location: Walnut Creek, CA

Join Date: Jan 2014
Posts: 2,707


Unfortunately, it looks like ploidy was estimated wrong; I don't think anything has a ploidy of 3. If the peak calls are correct, it's probably a highly polymorphic tetraploid, or a highly repetitive diploid or haploid. The full genome size would be approximately:

1*100786874 + 2*913749 + 4*41304813 + 8*1199939 + 10*138959 ...etc

Yielding 278822726, and so including all the peaks it would be around 285812914 as the program indicated. However the multipliers for the various peaks are hard to determine. Generally, the multiplier is the the peak's center divided by the first peak's center, so the first multiplier is 31/31=1, the second is 65/31=2. But the third is 112/31=3.6, which is hard to tell if it's 3 or 4 (they are all integers). It's probably 4.

Thus the genome is around 285Mbp all told, but it's hard to figure out the ploidy which makes it hard to figure out the haploid size. If it's a haploid then the genome is ~285MB. If it's tetraploid then you divide it by 4 and get 4 copies of a 71Mbp genome. For example, a human cell has around 6 Gbp of DNA in it, but because humans are diploid, the haploid representation of the genome is around 3 Gbp long.

The number of unique kmers is not important. Fold coverage is the amount of times you have sequenced each base, on average. Haploid genome size is the size of the genome divided by the ploidy. Haploid fold coverage is the number of times you have sequenced each base in the haploid representation, if the ploidy estimate was correct, which it isn't. Ploidy is the number of copies of each chromosome per cell. Het rate is an estimate of the fraction of the genome that is heterozygous (usually meaning a SNP); so 0.03413 indicates that 3.4% of the genome is heterozygous, meaning there is a SNP every 30 bases, which is very high. Percent repeat is the percent of the genome that occurs more than once as an exact copy; this number comes from analyzing the high-depth peaks, which come from repeat regions. However, if the ploidy estimate is off, then the het rate and percent repeat will also be off.

Usually the ploidy estimate works well for haploid and diploid organisms, but tetraploid and higher are difficult.

As for the post you pointed to, I was totally wrong about the formula, it should have been multiplication, not division. I will edit it.

Last edited by Brian Bushnell; 11-06-2015 at 10:41 AM.
Brian Bushnell is offline   Reply With Quote