dna-visualization

dna-visualization Project

This is a Quarkus port of 504.dna-visualization from SeBS: Serverless Benchmark Suite.

This application reads a DNA sequence from an input file and transforms it to a sequence of two-dimensional vectors using the Squiggle method.

Note that the input any character other than A, T, C, G is treated as invalid and is translated to a horizontal segment (i.e., a sequence of two (0.5, 0) vectors.)

Preparing Input Data

This application reads a DNA sequence file in FASTA format as input. This program uses JFASTA to read a DNA sequence from a FASTA file, so the file format should be acceptable by JFASTA.

The input files we tested are:

• The sample FASTA file published by EHT Zürich team
• DNA sequence of Methanocaldococcus jannaschii (ungzip after download)

The input file is assumed to be stored in Cloud Object Storage (COS). COS environment variable configuration is described in benchmarks/UsingCloudObjectStorage.md.

(FYI) Methanocaldococcus jannaschii is the first archaeon to be genome sequenced completely. Its data size is less than a half of the sample data published by ETH Zürich team.

Building and Running the Application

Build the project as described in this README.

Then, the application can run as a local HTTP server. To run the stand-alone Java version:

java -jar target/quarkus-app/quarkus-run.jar

To run the stand-alone native version:

target/dna-visualization-1.0.0-SNAPSHOT-runner

Sending a Request to the Application

This application receives the following parameters from POST data in JSON format:

Name	Value	Required?	Default	Default is customizable?
input_bucket	COS bucket to download input files	Y	(None)	Y
output_bucket	COS bucket to upload output files	Y	(None)	Y
key	COS object key of the input/output file	Y	(None)	Y
debug	Flag if output is uploaded to COS	N	false	N

For example:

curl http://localhost:8080/dna-visualization \
     -X POST \
     -H 'Content-Type: application/json' \
     -d '{"input_bucket":"MyInputBucket", \
          "output_bucket":"MyOutputBucket", \
          "key":"FASTAfiles/NewSequence.fasta"}

The above example downloads a FASTA file FASTAfiles/NewSequence.fasta from a COS bucket MyInputBucket, transforms the DNA sequence to a two-dimensional plot using the Squiggle method, but does not upload the output to a COS bucket.

Note that uploading a file to COS can take much longer than downloading a file and transforming the DNA sequence. Therefore, skipping uploading the file is recommended for evaluation of performance.

To send a request to a Knative eventing service:

curl http://<broker-endpoint>:<port>/ \
     -v \
     -X POST \
     -H 'Ce-Id: 1234' \
     -H 'Ce-Source: curl' \
     -H 'Ce-Specversion: 1.0' \
     -H 'Ce-Type: dna-visualization' \
     -H 'Content-Type: application/json' \
     -d '{"input_bucket":"MyInputBucket", \
          "output_bucket":"MyOutputBucket", \
          "key":"FASTAfiles/NewSequence.fasta", \
          "debug":"true"}'

This request downloads a FASTA file FASTAfiles/NewSequence.fasta from a COS bucket MyInputBucket, transforms the DNA sequence to a two-dimensional plot using the Squiggle method, and uploads the output to a COS bucket MyOutputBucket with the same object key because debug parameter is set to true.

Customizing the Default Value of Input Parameters

The default values of the three parameters can be customized by using environment variables or by modifying resources/application.properties. The following table describes this.

Name	Environment Variable	Key in application.properties
input_bucket	KNATIVEBENCH_DNA_VIS_INPUT_BUCKET	knativebench.dna-vis.input_bucket
output_bucket	KNATIVEBENCH_DNA_VIS_OUTPUT_BUCKET	knativebench.dna-vis.output_bucket
key	KNATIVEBENCH_DNA_VIS_KEY	knativebench.dna-vis.key