Varied items such as a pop music video from OK Go, Archimedes' "The Methods of Mechanical Theorems," and the Crop Trust's seed database have all been stored in DNA.
"For sheer density of information storage, DNA could be orders of magnitude beyond silicon — perfect for long-term archiving," Nature News says. But to make DNA data storage a more viable option, Nature News adds that issues of error-correction and random-access reading, among others, need to be contended with.
During DNA synthesis and sequencing, errors tend to crop up. To combat that, George Church's lab at Harvard Medical School has been relying on the redundancy of multiple copies to fix errors while researchers at the European Bioinformatics Institute, including Nick Goldman, have been taking a more complex approach using base three and overlapping, 100-base-long stretches of DNA. Even with these four versions of each 25-base piece, Nature News notes that they lost two of those sequences in a recent test.
At the same time, Microsoft researchers and others are working on way to access only the bits of DNA storing the data the user is interested in and not the entire DNA sample, Nature News adds. For instance, University of Illinois at Urbana–Champaign researchers are appending tags at the ends of DNA encoding data so they can be readily plucked out using PCR or CRISPR/Cas9.