Table 7. Computer Algorithms to Predict Protein Solubility Based on Protein Sequence.
| year | name | database | accuracy | Matthew’s correlation coefficient | ref |
|---|---|---|---|---|---|
| 2021 | GraphSol | 2737 | 0.78 | (456) | |
| 2019 | ProGan | 3148 | (457) | ||
| 2018 | DeepSol | 129643 | 0.77 | 0.55 | (96) |
| 2018 | PaRSnIP | 129643 | 0.74 | 0.48 | (458) |
| 2017 | Protein-Sol | 2359 | 0.90a | (93) | |
| 2016 | PON-Sol | 1080 | 0.49 | (459) | |
| 2014 | Niu-Li | 9500 | 0.88 | 0.76 | (753) |
| 2013 | ESPRESSO | 4922 | 0.68 | 0.42 | (460) |
| 2012 | PROSO II | 82299 | 0.75 | 0.39 | (461) |
| 2012 | CCSOL | 3043 | 0.74 | (462, 463) | |
| 2012 | SCM | 957/16902/82299 | 0.72 | (464) | |
| 2009 | Samak-Wang | 3173 | 0.8 | (465) | |
| 2009 | SOLpro | 17408 | 0.74 | 0.49 | (92) |
| 2007 | PROSO | 14200 | 0.72 | 0.44 | (12) |
| 1991 | PRSP | 81 | 0.88 | (466) |
a
Accuracy at 58% solubility prediction threshold.