Table 2.
Count | YLoc+ | DeepLoc 1.0 β | Fuel-mLoc | LAProtT5 | DeepLoc 2.0 | ||
---|---|---|---|---|---|---|---|
Animalα | Euk γ, θ | ESM1b | ProtT5 | ||||
Type | Multi | Single | Multi | Single | Multi | Multi | |
Pred. Num. Labels (Actual: 1.22) | 1.44 | 0.89 | 1.00 | 0.94 | 1.15 | 1.21 | |
Accuracy | 1717 | 0.23 | 0.37 | 0.38 | 0.45 | 0.34 | 0.39 |
Jaccard | 1717 | 0.41 | 0.42 | 0.46 | 0.52 | 0.48 | 0.53 |
MicroF1 | 1717 | 0.51 | 0.46 | 0.52 | 0.56 | 0.57 | 0.60 |
MacroF1 | 1717 | 0.34 | 0.35 | 0.39 | 0.43 | 0.44 | 0.46 |
MCC per location (↑ is better) | |||||||
Cytoplasm | 562 | 0.14 | 0.23 | 0.23 | 0.33 | 0.29 | 0.36 |
Nucleus | 893 | 0.20 | 0.28 | 0.41 | 0.45 | 0.41 | 0.44 |
Cell membrane | 287 | 0.20 | 0.23 | 0.32 | 0.30 | 0.34 | 0.36 |
Mitochondrion | 196 | 0.37 | 0.39 | 0.33 | 0.59 | 0.60 | 0.56 |
Endoplasmic reticulum | 77 | 0.12 | 0.23 | 0.14 | 0.22 | 0.20 | 0.17 |
Golgi apparatus | 86 | 0.08 | 0.10 | 0.24 | 0.26 | 0.17 | 0.31 |
Bold values indicate the best score
α = GO-terms were not used
β = Retrained on the new CV dataset
γ = using local implementation
θ = using reduced ProSeq database