Table 1.

The effect of tile-size on Geoseq results

Tile size	Sequences retrieved	System time (in seconds)
15	7049	0.52
18	1135	0.54
20	74	0.51
21	32	0.52
25	14	0.52
28	14	0.48
30	4	0.5

The tile size determines the number of reads selected for further analysis, which determines sensitivity, specificity and speed. Reads were retrieved using Geoseq with different tile sizes for a 1158-nt-long terminal exon from a variant of ASB10 (chr7:150883831-150884989). There is a sharp transition in the number of reads retrieved at a tile-size of 20; at lower tile sizes a majority of the retrieved reads map to multiple locations on the genome. We picked 25 as an optimum tile size based on an expected average of 1 variant per 100 nt in the genome. The time to align the retrieved reads against the exons grows linearly in the number of sequences.