G4 stability of PQSs in the human nuclear genome

Interestingly, against the prior considerations, 50.96% of PQSs do not appear to form stable genomic G4 structures (Fig. 3A, see below for the mapping details), with the remaining 49.04% accounting for the 65.56% of all the G4-seq-observed G4s (525,890 under K⁺/Na⁺ contrasting condition⁸). The rest of the observed G4s were those of non-canonical nature⁸, such as with bulges³⁶ and shorter-than-3-nt span for guanines in G-tracts^37,38. In other words, the usage of the extended PQS motif with 12-nt maximum loop length, though capturing 65.56% of the experimentally observed G4s in the human genome, contained inherent 50.96% false positives (see above). However, the use of the more conservative PQS definition with 7-nt maximum loop-size captured only 36.86% of observed G4s (instead of 65.56%), but with not-much-different, 46.37% fraction of false positives. Therefore, the preferred general motif to hunt for G4s is PQS with 12-nt maximum loop length. However, a human sequence conforming that motif would still have only a 49.04% chance of forming a stable genomic G4.