"BioModels is a repository of mathematical models of biological and biomedical systems. It hosts a vast selection of existing literature-based physiologically and pharmaceutically relevant mechanistic models in standard formats. Our mission is to provide the systems modelling community with reproducible, high-quality, freely-accessible models published in the scientific literature."
Another fascinating thing I just discovered existed. So these are mathematical models, not neural network models, as we are used to hearing the word "model" used (at least if you follow my stuff). So they are things like population models, metabolic networks, petri nets, ordinary differential equation models, delayed differential equation models, stochastic differential equation models, partial differential equation models, differential algebraic equation models, boolean models, cellular Potts models, agent-based models, pharmacodynamics models, constraint-based models, steady-state models, rule-based models, protein-protein interaction networks, Markov chains, finite element spatial models, and physiologically based pharmacokinetic models.
All the models are supposed to be in a data format called Systems Biology Markup Language (SBML). In addition to SBML, some are in other formats like MATLAB/Octave, COMBINE archive (whatever that is), MorpheusML (whatever that is), CompuCell3DML (whatever that is), as well as programming languages Python, R, and Mathematica, and you can filter on these when you do searches. They also promise a human readable summary of each model in PDF format.
You can also filter by organism, such as Homo sapiens, Saccharomyces cerevisiae, Mus musculus, Escherichia coli, Drosophila melanogaster (fruit fly), and you can filter on larger categories like Mammalia, Vertebrata, Eukaryota, Chordata, etc.
You can also filter by disease, such as Alzheimer's, Covid-19, Diabetes Mellitus, Cancer, Parkinson's, Osteoarthritis, etc.
They have a filter for "GO", which means gene onotology. The Gene Ontology project aims to come up with the same or similar names for genes that have the same function across all species. Here the "GO" filter has such categories as cellular metabolic processes, cytoplasm, nucleus, translation (going from DNA to proteins), extracellular regions, cytosol, protein catabolic process, cell death, endoplasmic reticulum, mitochondrion, lysosomes, Golgi apparatus, peroxisome, cell growth, blood coagulation, protein phosphorylation, ATP hydrolysis, and so on (very long list actually -- hundreds of items).
You can filter on "UniProt", which is a "Universal Protein Database". Proteins have names like Pyruvate kinase PKLR, Cytoplasmic aconitate hydratase, Glucose-6-phosphate isomerase, Hydroxymethylglutaryl-CoA synthase, and Fructose-bisphosphate aldolase A. Well, those all have -ase names, indicating they are enzymes, but not all proteins are enzymes, of course, so the database has names like ATP-binding cassette sub-family A member 1 as well.
You can filter on "ChEBI", which stands for "Chemical Entities of Biological Interest". Here you can search for things like ADP, ATP, glycerol, phosphoenolpyruvate, acetaldehyde, aldehydo-N-acetyl-D-glucosamine, hydrogen peroxide, glycine, ethanol, NADH, etc.
The last thing they have that you can filter on is Ensembl. I actually told you all about Ensembl before, when 2.5 years ago I told you about the "mitochondria calcium channel mystery". Our mitochondria calcium channel is made of 3 proteins, but fungi have only 2 of them. With some clever research it was determined that the ancestor of both us and fungi had all 3 proteins, but fungi lost one of them. The researchers used 1,156 eukaryotic genomes to do this, but they didn't sequence those 1,156 eukaryotic genomes themselves -- they just download them from the Ensembl Project website. I thought it was amazing something so far in the distant evolutionary past was possible to determine, and all just by downloading data from this already existing database. The Ensembl Project was created in tandem with the Human Genome Project in the 90s. Lookups into this database go by actual gene name, which are those cryptic all-caps-and-numbers designations, such as NFKB2, NANOGP1, POU5F1, QSOX2, CASP8, etc.