Skip to main content

Table 8 Practical results for different alphabets – Quality estimations

From: Optimal neighborhood indexing for protein similarity search

alphabets number of positions validating Stage 1 and Stage 2 practical selectivity number of detected alignments practical sensitivity
Σ20 × Σ20 2.14 * 106 1.35 * 10-3 650 (all) 1
Σ20 × Σ16 1.39 * 106 0.88 * 10-3 650 (all) 1
Σ20 × Σ16 0.98 * 106 0.62 * 10-3 650 (all) 1
Σ20 × Σ8 0.62 * 106 0.39 * 10-3 650 (all) 1
Σ20 × Σ4 3.14 * 106 1.98 * 10-3 650 (all) 1
Σ20 × Σ2 2.93 * 106 1.85 * 10-3 650 (all) 1
  1. Similarity search results obtained on reduced alphabets. The number of positions tested (validating Stage 1 only and independent from the chosen alphabet) is 1.59 * 109. The practical selectivity is computed dividing the number of positions validating both Stage 1 and Stage 2 by the number of positions tested.