Chronologer Retention Time model predictor #761
Conversation
| /// The base sequence is the sequence without modifications and the full peptide sequence is the sequence with modifications. | ||
| /// The model is intended to be used with sequences of length 50 or less, and only supports modifications present in the Chronologer dictionary. | ||
| /// Sequences not supported by chronologer will return null. | ||
| /// </summary> |
There was a problem hiding this comment.
add example full sequence to tensor
…to Chronologer
| } | ||
|
|
||
| if (baseSequence.Length <= 50) //Chronologer only takes sequences of length 50 or less | ||
| { |
There was a problem hiding this comment.
If statement is unnecessary due to above check. that will give you coverage of lines 222-223
mzLib/Proteomics/RetentionTimePrediction/ChronologerModel/ChronologerEstimator.cs
Show resolved
Hide resolved
| { ('S', "Phosphorylation on S"), 25 }, //'Phosphoserine | ||
| { ('T', "Phosphorylation on T"), 26 }, //'Phosphothreonine | ||
| { ('Y', "Phosphorylation on Y"), 27 }, //'Phosphotyrosine | ||
| { ('K', "Accetylation on K"), 28 }, //acetylation |
There was a problem hiding this comment.
What about alternative spellings? Akin to the issue from Zhuoxin's two phosph on S bars in her presentation. Unimod may say: Phospho on S where we may say Phosphorylation on S
There was a problem hiding this comment.
Why are we downgrading easy.common?
| namespace Omics.Fragmentation.Peptide | ||
| { | ||
| { |
| catch | ||
| { | ||
| Console.WriteLine(new Exception("This device does not support CUDA, using CPU")); | ||
| device = new torch.Device(DeviceType.CPU); |
There was a problem hiding this comment.
Is there a way to do this check without catching an exception?
| { | ||
| preds[outputIndex] = predictionHolder[outputIndex]; | ||
| }); | ||
|
|
There was a problem hiding this comment.
what does the "preds" array do here? It doesn't look like it's used
| var baseSeq = baseSequences[i]; | ||
| var fullSeq = fullSequences[i]; | ||
|
|
||
| var (tensor, compatible) = Tensorize(baseSeq, fullSeq, out bool chronologerCompatible); |
There was a problem hiding this comment.
what is the difference between compatible and chronologerCompatible? Why is one checked but not the other?
This PR adds the Chronologer model and the weights from the article.
Proteomics csporj file now contains a conditional installing statement for the operating system, this is to enable CUDA for compatible OS