Talk:Fabry:Sequence-based mutation analysis

Praise

• Nice tables
• Nice visualizations of the mutated residues. Is this pymol?

I don't think so. They were generated by HOPE.

• Plenty of physicochemical properties.

Criticism

• Does the secondary structure prediction refer to the sequence before or after the mutation?

--> Since we used the files from Task 3 (see Journal), it is from the native structure

• What are your reasons for taking into account the substitution score of PAM1 and PAM250? Is the evolutionary distance not too short and too far, respectively?

--> We were said to do that. At least that's how I understood it

• I think the substitution scores of the three matrices are not directly comparable. PAM1 should have lower scores than PAM250 relative to the maximum score.

--> We díd compare them relatively not the absolute values

• What do the entries in the PSSM table mean? It is obviously not the substitution score. In case of P40S: is P(P|40) = 0.81 or P(S|40) = 0.81?

It's the "weighted observed percentages rounded down" value as shown here. As I understood, it means that, e.g. for P40S, in 81% of the sequences, the wild type is conserved whereas in 2% of the sequences, the mutation (P → S) is observed.

• You classify the features either as disease causing or not disease causing. But what if a feature is in between?

--> There are more severe forms of the disease and less severe forms, but either you are ill or not... What else is there in between?

• You are weighting each feature equally. For instance PAM1, PAM25, and BLOSUM62 are treated equally. The three secondary structure prediction are of course not independent.

--> but sometimes different. A weighting function would have actually been a good idea, but since we are not done yet, it would have made no sense to develop one.