ASMscan-BiLSTM

Bidirectional LSTM model for detection of amyloid signaling motifs (ASMs).

Installation

pip install asmscan-bilstm

Usage

from asmscan.bilstm import BiLSTM

aa_seqs = [
    "MEGRASGSARIYQAGGDQYIEESDGYRADG",
    "VSLRAGAHDGGRIYQAVGDQYIYEDGGRAGASLREDGYQRIYAGASIYEA",
    "HASGHGRVFQSAGDQHITEHAGHDG"
]

model = BiLSTM()
pred, frags = model.predict(aa_seqs)

The predict() method of the BiLSTM model position-wise averages the predictions of the 6 cross-validation models, and then applies max-pooling over the sequence length. Sequences are processed using a window of size 40 and a step of 1.

References

ASMscan-BiLSTM model is part of the ASMscan project:

• Not yet published.

ASMscan project is an extension of the ASMs analysis conducted with the PCFG-CM model:

• W. Dyrka, M. Gąsior-Głogowska, M. Szefczyk, N. Szulc, "Searching for universal model of amyloid signaling motifs using probabilistic context-free grammars", BMC Bioinformatics, 22, 222, 2021.

• W. Dyrka, M. Pyzik, F. Coste, H. Talibart, "Estimating probabilistic context-free grammars for proteins using contact map constraints", PeerJ, 7, e6559, 2019.