About Prediction Levels
Level I: This is first level where a query sequence is being predicted as single localization or dual localization
Level II: This level will run leve1 first followed by level2 where a query sequence is first predicted as single localization or dual localization followed by classifying single localization into 10 classes. 'Vacuole', 'Cytoplasm', 'Golgi', 'Endoplasmic', 'Memebrane', 'Secreted', 'Cell', 'Mitochondria', 'Plastid','Nucleus'.
Level III: This level will execute level1 and level2 followed by classifying dual localization in 6 classes. 'Cell-memebrane & Membrane', 'Cytoplasm & Membrane','Endoplasmic reticulum & Membrane','Golgi & Membrane','Nucleus & Chromosome','Nucleus & Cytoplasm'.
level IV: This phase will execute level1, level2 and level3 followed by classifying Memebrane predicted in level2 into single-pass or multi-pass.
About Prediction Strategy
There are two prediction strategy available in RSLpred-2.0 [Fast, Sensitive]. There is an icon with general information about the tool and a brief explanation on how it works.
The 'Fast' approach model is designed for swift predictions. In its fast mode, the model utilizes a DPCP feature, a composite of Dipeptide composition (400), amino acid pair count (400), and Schneider-Wrede values for each amino acid pair (400). The resulting DPCP feature vector has a total size of 1200. The efficiency of this approach is attributed to its smaller vector size, facilitating rapid processing. This makes the 'Fast' approach particularly beneficial for annotating a large volume of proteins with speed and accuracy.
The 'Sensitive' approach model excels in delivering heightened sensitivity in predictions, albeit with a trade-off of increased computation time. In its sensitive mode, the model employs the TPC (tripeptide composition) feature, generating an extensive 8000-length vector. The extended computation time is a result of the larger vector size, making this approach particularly sensitive. However, the 'Sensitive' approach proves valuable for annotating a smaller number of proteins with a focus on high-quality predictions.
Error in Job Submission
Error.