Detection of the HERV transcriptome.

aNumber of distinct genomic HERV loci included in HERV database HERV-gDB3. The database contains 6 HERV families with unequal input. The search for distinct elements belonging to each family is performed by systematic BLAST genome coverage, allowing a maximum 20% divergence with prototype elements.
bNumber of distinct genomic HERV loci present in the chip. Each element of the database is processed through home-made EDA+ algorithm to find probes that match optimal hybridization criteria. The candidate probes are then checked against the entire human genome (NCBI 36/hg18) using the KASH algorithm to control their cross-hybridizing ability and non-specific sequences are removed. Probes are ultimately assembled into probesets to discriminate individual genomic HERV sequences. Differences between database and chip mark the success in designing HERV-specific probes and probesets. For clarity, the probeset content is not detailed.
cHERV transcriptome results: number of active elements in all tissues tested. After the experiments were normalized using the COMBAT method and an arbitrary positive threshold was applied (value = 100), elements that are active in at least one tissue are enumerated.
dSubsets of complete or partial proviruses.
eOne element can be composed of several probesets.