In case of single restriction enzyme digestion, both anchors can potentially ligate to both MseI generated sticky end but only AB or BA containing fragments can be used for sequencing. This might give the suggestion that 50% of the VIDISCA products are ineffective as they contain the same adaptor. However, the fragments containing 2 different primers are preferentially amplified in the PCR, since an AA or BB fragment has a disadvantage that 59 and 39 ends anneal to each other which interferes with primer annealing. We definitely observed the higher chance of amplification of several genome segments when only one restriction site is used. Remarkably high genome coverage was noted in several samples, a coverage which could never be achieved in case two restriction enzymes were used in amplifications. Other groups have used high throughput sequencing for virus discovery as well. In one paper the viral community in an Antarctic lake was described. Lopez-Bueno et al. collected water in spring and late summer from a fresh water lake in Antarctica and used high throughput sequencing to study the viral community in a location hardly visited by larger eukaryotes. For the first time a large amount of sequence data was retrieved from this isolated place which led to the identification of at least 12 viral families of which two are claimed to represent new families. Their results show the enormous possibilities for virus discovery and high throughput sequencing. The authors also address a large amount of unknown sequences present in their data set. We also observed the presence of unknown sequences within our data set. It could be that these sequences are derived from yet unknown viruses, or it could be that the sequences are part of a genomic sequence from a known organism, e.g. a bacterium of which not the complete genomic sequence is present in the Genbank databases. Thus care should be taken to assign sequences as potentially viral, since so many organisms have not been fully sequenced. There are several advantages of high throughput sequencing in comparison to BigDye terminator sequencing. First of all, with high throughput sequencing and pooling of samples that carry their own recognition sequence the VIDISCA cost per sample is reduced, since selective VIDISCA-PCR, metaphor agarose gel AZ 960 visualization, purification of fragments from gel, TA cloning, colony PCR and subsequent BigDye sequencing can all be omitted. Secondly, the amount of sequence data received from a single sample is higher than what can be achieved in standard VIDSCA, thus increasing the chances of identifying an unknown virus. This method opens new opportunities for virus discovery, not only in respiratory samples of undiagnosed respiratory infection, but also in diseases such as Amyotrophic lateral sclerosis, Kawasaki disease and Multiple sclerosis. For these syndromes a viral pathogen has been suggested but could not be confirmed so far.