Developing eThread Pipeline Using SAGA-Pilot Abstraction for Large-Scale Structural Bioinformatics

Algorithm 3

Proposed algorithm combining task-level parallelism and dynamic scheduling for eThread on EC2.

Data:  𝑁 protein gene sequences

Result: Gene annotation and tertiary structure

forall the 𝑁  sequences do

   read sequence

   forall the 10 threading tools do

        forall the domain, chain do

     while task to run do

       estimate tts, memory footprint

       map available resource

       if pre-processing then

          do pre-processing

       end

       do main processing

       do post-processing

     end

       end

       write about

   end

end

for 𝑁  sequences do in parallel

   /∗ now meta-analysis step using all

       outputs from 10 threading tools

       ∗/

   read all outputs

   do meta-analysis

end