A maximum gap value ∼240 meV is obtained around ±4% strain.
The heuristic solutions obtained are within a maximum gap of 13% from the LP relaxation.
A binding p-value was then determined for each genomic position by Wilcoxon rank sum test and binding sites were generated from those more significant than specified thresholds with a maximum gap of 500 and minimum run of 350.
However, there were 23 gaps of >2 Mbp in E, with a maximum gap of 19.7 Mbp in G4, adding to a total distance of 103.0 Mbp, almost half of the distance covered by T×E or T1E.
For PASS, a maximum gap of 8 bases was allowed with a gap open and extension penalty of 1.
A maximum gap size is set to the longest gap seen in the training data plus 20, for computational efficiency.
Gaps were rare in T×E and T1E, four in total for each map with a maximum gap of 2.9 Mbp.
The average density of anchor markers was of 7.3 cM/marker with a maximum gap of 22 cM (in PG6) for Prunus and 8.0 cM/marker and a maximum gap of 32 cM (in FG3) for Fragaria.
A maximum gap size was enforced between scans; for the 65% missing scan pattern, the maximum spacing permitted between scans was 55μm.
