Protein structure search and local structure characterization

BMC Bioinformatics

Table 1 Relationship between the number of clusters found and the number of SOM map units used

SOM map size	Number of clusters	SOM map size	Number of clusters
10 × 10	6	110 × 110	24
20 × 20	9	120 × 120	19
30 × 30	10	130 × 130	21
40 × 40	12	140 × 140	22
50 × 50	15	150 × 150	18
60 × 60	13	160 × 160	15
70 × 70	14	170 × 170	21
80 × 80	18	180 × 180	18
90 × 90	18	190 × 190	18
100 × 100	20	200 × 200	18

Our analysis determined that among the number of clusters that maximized the BIC, 18 clusters occurred most frequently. Thus, we assigned 18 letters to our alphabet.

ISSN: 1471-2105