nerdalize/psspred repository overview

⁠PSSPred

PSSPred is a a multiple neural network training program for protein secondary structure prediction

This repository contains the Docker instructions for creating a Docker Image.

BEWARE! This image is something like 8GB in size

⁠Usage:

PSSPred is a perl program which expects an input file as it's only input argument and will output a file: seq.dat.ss. That doesn't really work well for batch jobs so this Docker Image is constructed to accept a BASE64 encoded single hemoglobin sequence as the input argument like so:

$ cat sequence.fa | base64 | xargs docker run nerdalize/psspred 

What this line does:

• cat reads the input of sequence.fa and prints it to stdout
• base64 base64-encodes that string, this is important, as otherwise the newlines will break things
• xargs takes that pipe, and turns it into an argument
• docker run nerdalize/psspred runs the docker image essentially with that argument, like: docker run nerdalize/psspred BASE64STRING

In the run.sh script we decode that input string again and write it to the inputfile.

⁠results

PSSpred.pl will output a couple of files in /workdir/. Amongst others there will be seq.dat.ss (the usefull result).

⁠ToDo:

(exercise for the reader)

• do something usefull with the output.