MetalNet is designed for detecting metal binding sites with coevolution data. And metal binding sites and metal type will be offered if predicted.

what you need:

1. msa file ($gene.msa)

2. coevolution profile ($gene.pair)

#$gene: input file name, such as 8f4a248217074d25a8bf3c71391e0bc9.msa

cd demo/input
bash ../../trim_coevolution.sh 8f4a248217074d25a8bf3c71391e0bc9
python ../../predict.py 8f4a248217074d25a8bf3c71391e0bc9

$gene.cnn: coevolved CHED pairs with predicted possibility

$gene.dat: prediction result after graph-based filter

$gene.gv.eps: visualization of predicted network cluster

Could be downloaded from GREMLIN server or calculated locally.

(reference instruction:)

jackhmmer -E 1E-20 -N 8 -A $gene.sto $gene.fasta {sequence_database}

sto2fas.pl $gene.sto > $gene.fas

hhfilter -i $gene.fas -id 90 -cov 75 -o $gene.msa

gremlin_cpp -i $gene.msa -o $gene.pair