Table 1 Running time and peak memory usage of all methods in searching GNPS against DNP.

From: MolDiscovery: learning mass spectrometry fragmentation of small molecules

Method

Searching time (d-h:m:s)

Searching peak memory (GB)

Preprocessing time (d-h:m:s)

Preprocessing peak memory (GB)

molDiscovery

00:29:04

0.15

10:40:25

16.38

CFM-ID

00:59:37

0.012

157-08:28:08

3.839

Dereplicator+

04:28:10

19.98

N/A

N/A

MAGMa+

2-20:34:37

0.51

N/A

N/A

MetFrag

12-08:57:24

4.01

N/A

N/A

CSI:FingerID

29-09:35:05

19.68

N/A

N/A

  1. Note that CFM-ID running time and memory usage does not include filtering database candidates by precursor m/z. All running times are only for the spectra for which the method completed successfully (exit code 0). For CSI:FingerID 423 spectra either failed with a "feasibility" error or required more than 20 GB of memory, and are not included in values in this table. Note that given a small molecule structure database, molDiscovery and CFM-ID only need to preprocess the database for once. For searching very large spectral datasets, they are usually more efficient than other methods.
  2. N/A not applicable.
  3. All reported running times are on a single CPU.
Back to article page