iMLP

iMLP predicts internal matrix targeting-like sequences (iMTS-Ls) in protein sequences and generates propensity profiles to inform mitochondrial protein targeting from the cytosol into mitochondria, taking into account structural similarity to mitochondrial targeting sequences (MTSs).

Key Features:

• Target prediction: Predicts internal matrix targeting-like sequences (iMTS-Ls) within protein sequences.
• Propensity profiles: Produces iMTS-L propensity profiles for input protein sequences.
• Deep learning algorithm: Implements a recurrent neural network (RNN)–based deep learning model to perform predictions.
• Biological context: Accounts for structural features shared between iMTS-Ls and classical mitochondrial targeting sequences (MTSs).
• Performance: Reports improved speed and efficiency compared to existing approaches.

Scientific Applications:

• Mitochondrial protein targeting studies: Identification of iMTS-Ls to investigate mechanisms of protein import from the cytosol into mitochondria.
• Protein annotation: Annotation of potential internal targeting signals in proteomes for mitochondrial localization analyses.
• Comparative sequence analysis: Comparative study of iMTS-Ls and classical MTSs to explore sequence and structural determinants of targeting.

Methodology:

Uses a deep learning approach employing a recurrent neural network (RNN) trained to predict iMTS-L propensity profiles for protein sequences.