Mycoplasma contamination in the 1000 Genomes Project (BioData Mining)

Background: In silco Biology is increasingly important and is often based on public data. While the problem of contamination is well recognised in microbiology labs the corresponding problem of database corruption has received less attention. Results: Mapping 50 billion next generation DNA sequences from The Thousand Genome Project against published genomes reveals many that match one or more Mycoplasma but are not included in the reference human genome GRCh37.p5. Many of these are of low quality but NCBI BLAST searches confirm some high quality, high entropy sequences match Mycoplasma but no human sequences. Conclusions: It appears at least 7% of 1000G samples are contaminated.