How to select mutation sites for wet experiment

[lsy-199612]

I conducted the saturated mutation prediction of the whole sequence and obtained the mutational effect and Predict probabilities. Is there any recommended screening threshold

Position,Mutation,Effect,Probability 1,A1A,0.0,0.026617998257279396 1,A1C,-2.8358705043792725,0.001561609678901732 1,A1D,0.43181079626083374,0.04099288210272789 1,A1E,0.5467181205749512,0.04598456248641014 1,A1F,-0.06809435039758682,0.02486579678952694 1,A1G,0.039064086973667145,0.027678383514285088 1,A1H,-1.6565617322921753,0.005078550428152084 1,A1I,-0.4099830687046051,0.01766533963382244 1,A1K,0.6258731484413147,0.04977240785956383 1,A1L,0.8428149223327637,0.061830855906009674 1,A1M,2.663308620452881,0.3817997872829437

[LTEnjoy]

Hi,

The higher the mutational effect score is, the better our model thinks such mutation should be. You can sort these rows by the mutational effect on a descending order. Then you could choose like top 10 mutations and conduct wet experiment. We have many wet experiments reported in our lates paper SaprotHub, see https://doi.org/10.1101/2024.05.24.595648. You could also use our colab server to make zero-shot prediction, without any coding, see https://colab.research.google.com/github/westlake-repl/SaprotHub/blob/main/colab/SaprotHub_v2.ipynb :)

[lsy-199612]

Hi,

The higher the mutational effect score is, the better our model thinks such mutation should be. You can sort these rows by the mutational effect on a descending order. Then you could choose like top 10 mutations and conduct wet experiment. We have many wet experiments reported in our lates paper SaprotHub, see https://doi.org/10.1101/2024.05.24.595648. You could also use our colab server to make zero-shot prediction, without any coding, see https://colab.research.google.com/github/westlake-repl/SaprotHub/blob/main/colab/SaprotHub_v2.ipynb :)

Is the range of the mutational effect score from 100 to -100? My highest prediction score is +6. Does this mean that the mutation prediction result is not very good

[LTEnjoy]

I think theoretically the range of the mutational effect score is from -∞ to +∞. You could check the definition of mutational score here. In my opinion, I think +6 is good enough.