During the last years, virtual screening(VS) methods have been
established as an alternative to high-throughput screening(HTS)
to discover potential lead structures for new therapeutic targets.
Considering information about the three-dimensional structure
of a protein, virtual screening methods search large molecule
databases to select promising ligands and to predict their interaction
with the protein.
 |
 |
| HTS |
VS |
Fig 1. High-Throughput Screening vs. Virtual
Screening |
- HTS of 400,000 compounds from a corporate collection
? 300 hits < 300 ?M = 0.021%
hit rate (many violate Lipinski rules)
- VS of 235,000 commercially available compounds, using DOCK 3.5
? 365 high-scoring molecules, 127
with IC50 < 100 킡 = 34.8% hit rate (hits are more drug-like)
Table 1. Hit rates from High-Throughput and Docking Screens
against PTP1B
| Technique |
Compds tested |
Hit with IC50 < 100 킡 |
Hits with IC50 < 10 킡 |
Hit rate(%) |
| HTS |
400000 |
85 |
6 |
0.021 |
| docking |
365 |
127 |
21 |
34.8 |
* J. Med. Chem. 2002, 45,
2213-222
Table 2. Comparison of HTS to VS
| |
Traditional Screening |
Virtual Screening |
| Probability to Clinical Test |
1/5000* |
1/10 - 1/100 |
| Time to Discovery(Year) |
3 - 5 **
|
1 - 2 |
Cost (M$)
|
10 - 120 * |
< 3 |
| Technology |
Trial & Error, Intuition
|
Rational |
Licensing Out (Year)
|
5 - 10
|
2 - 3 |
* Drug Discovery Filtering Out
Failures Early in the Game - June 5, 2000 C&EN
** Re-Inventing Drug Discovery -
Andersen Consulting (1998)
|