| AMDIS Benchmarks AMDIS is a very CPU and memory intensive program. The following procedures compare the times
for deconvolution and library search under different conditions. - AMDIS test file: 50p0830d; pure standards: low noise; peak shape OK
- EPA method 524.2 purge and trap GC-MS analysis of volatile organic compounds in water (60 volatile compounds).
- target library: NISTEPA.MSL (1112 compounds)
- component width: 20
- VL=very low; L=LOW; M=midlle; H=high; VH=very high
-
WIN98; Athlon 1700Mhz; 512 MB RAM; IBM ICSL HD
Results: Adjacent (neighboring) peak subtraction (low to high) has no impact in low res mode (see end of this table) - but in high res mode the number of target analytes will increase. Resolution (low to high) increases number of target analytes and components. Sensitivity (low to high) increases number of target analytes and components. Shape requirements (low to high) has no impact in low res
mode, but will decrease the number of targets and components in high res mode. Recommendation: 1. Start with a peak subtraction of 2, shape req. high, low resolution and very low sensitivity 2. Change sensitivity to medium 3. Change resolution to medium After all steps do a target library search or NIST search and compare your results. All these factors strongly depend on the kind of your data file (noise, peak overlap etc.) Benchmark Table
Peak subtract. | Resolution | Sensitivity | Shape Req. | Targets | Components | Time[s] | 2
| L | VL | L | 77 | 95 | 2.8 | 2 | L | L
| L | 91 | 117 | 3 | 2 | L | M | L | 100 | 153 | 4 | 2 | L | H | L |
102 | 196 | 4 | 2 | L | VH | L | 126 | 280 | 6,5 | | | | | | | | 2 |
L | VL | L | 77 | 95 | 2,8 | 2 | L | VL | M | 78 | 97 | 2,5 | 2 | L | VL | H | 79 | 95 | 2,5 | 2 | L | VH | L | 126 | 280 | 7 | 2 | L | VH | M | 129 | 277 | 6 | 2 | L | VH | H | 110 | 213 | 5 | | | | | | | | 2 | M
| VL | L | 159 | 167 | 6 | 2 | M | L | L | 194 | 233 | 9 | 2 | M | M |
L | 233 | 326 | 10 | 2 | M | H | L | 252 | 361 | 10 | 2 | M | VH | L | 307 |
559 | 15 | 2 | M | VH | M | 276 | 550 | 15 | 2 | M | VH | H | 229 | 476 | 13 |
| | | | | | | 2 | H | VL | L | 249 | 259 | 9 | 2 | H | L | L | 302 | 373 | 12 | 2 | H | M | L | 370 | 519 | 17 | 2 | H | H | L | 377 | 586 |
19 | 2 | H | VH | L | 432 | 878 | 25 |
2 | H | VH | M | 392 | 868 | 24 | 2 | H | VH | H | 306 | 775 | 20 | | |
| | | | | 2 | H | VL | M | 208 | 259 | 9 | 2 | H | M | M | 320 | 522
| 16 | 2 | H | VH | M | 392 | 868 | 24 | 2 | H | VL | H | 173 | 252 | 14 | 2 | H | M | H | 267 | 496 | 14 | 2 | H | VH | H | 306 | 775 | 20 | |
Impact of peak subtraction Low res mode: Trend of increasing peak subtraction has no effect High res mode: Increasing targets but constant components and increasing analysis time | N | L | VL | L | 74 | 95 | 2,5 | 1 | L | VL | L | 76 | 95 | 2,5 | 2 | L | VL | L | 77 | 95 | 2,8 | N | H | VH | H | 178 | 748 | 7 | 1 | H | VH | H | 219 | 777 | 12 | 2 | H | VH | H | 306 | 775 | 20 |
|
|