In StarDrop versions before v7.5, we have the P450 module and only P450 calculations were available, these are equivalent to this option in the Metabolism module:
In StarDrop versions above 7.5, this option would run at roughly the same speed as the P450 calculations in older versions, although there may be some differences due to the server implementation.
Predict Metabolism will typically run multiple enzyme simulations—up to five enzymes, depending on the output of the WhichEnzyme model. Each of these simulations is more complex than the P450 models, although there will be fewer sites of metabolism, so there’s a bit of a trade-off. But it’s not unreasonable to expect Predict Metabolism to be up to five times slower than Predict P450 Metabolism.
If you choose Generate Metabolic Pathway, this will run Predict Metabolism for several compounds. For example, if you consider Codeine:
To generate the second-generation metabolites, Predict Metabolism must be run on all of those compounds. It would not be unreasonable for this to be ~30 times slower than a ‘simple’ P450 calculation (6 metabolites and 5x more expensive). As this is a simple molecule – more complex molecules will take longer for each calculation and are likely to have more first-generation metabolites that each need to be predicted.
As for the number of jobs in the queue, there will be one job per potential site of metabolism for each compound (parent and first-generation metabolites). A big compound will have many potential sites of metabolism and potentially result in many first-generation metabolites, so this will lead to many jobs!
While the calculations are slower, they are very sophisticated and provide much more valuable information than the ‘simple’ P450 models.