The J. Craig Venter Genome Project

While this was make a great name for neo-hippie, alt-rock, acid-jazz band it is actually the subject of a recent PLoS Biology article (available here). Basically, the idea was to sequence the entire diploid genome of a single human (roque scientist J. Craig Veneter in this case) and compare it to the published version of the human genome.

This article represents a significant leap forward in the ability to understand the human genome. As humans we each generally have two copies of 22 of our chromosomes, and two sex chromosomes (XX = female), (XY=male). Most of the other genomic sequences have consisted of an ammalagmation of different individuals and generally been considered the genetic makeup of a single chromosome. This is the first attempt to sequence the entire genome and it was done over several years (basically required 32 million different sequence reads and several donations of biological material from Dr. Venter) and used the shotgun sequencing technology that Celera developed for the human genome reference sequence. It is a not a complete sequence as only 59% of the Y-chromosome was covered, the X-chromosome was at 95.2% and the autosomes at 98.3%. There were a few errors as well as described in the article

fraction of these putative variants resulted from sequence reads with variant base having reduced quality value (QV) scores, the presence of variants in homopolymer runs and erroneous base calls at the beginning and end of reads. The inclusion of these reads was important to the assembly process, and therefore we chose to perform post-assembly processing to filter these variants to reduce false positives while limiting false negatives fraction of these putative variants resulted from sequence reads with variant base having reduced quality value (QV) scores, the presence of variants in homopolymer runs and erroneous base calls at the beginning and end of reads. The inclusion of these reads was important to the assembly process, and therefore we chose to perform post-assembly processing to filter these variants to reduce false positives while limiting false negatives.

They then went ahead and corrected the errors to provide robust results. I found this latter part interesting considering some of the recent articles on the Neanderthal autosomal sequences that I mentioned earlier.

By far the most interesting aspects of this article are the results. They found more than 4.1 million DNA variants, composed of SNPs, Indels, block subsistutions, insertion/deletion events, etc. This has important implications for understanding human genetics as it may indicate that individuals have tremendous amounts of variation within each genome. While the cost of this is still too expensive to apply to human population genetics, in the next 10 years it may go down as the technology progresses, and may give greater insight into population subdivision.