AF | Home

Welcome to AFproject

a joint effort to benchmark, improve and simplify alignment-free sequence analysis

Summary of service

AFproject is a free service for objective performance comparison of alignment-free sequence comparison tools on different datasets.

This community developed web-service aims at simplyfing and standardizing the benchmarking of alignment-free tools. And for the users, the benchmarks provide a way to identify the most effective methods for the problem at hand.

Service content

AFproject reports performance assessment of alignment-free methods across 12 reference data sets, which are divied into 5 research applications.

#	Research application	Reference data set	Sequence type
1	Regulatory Sequences	Cis-regulatory modules (CRM)	non-coding DNA
2	Protein Sequence Classification	Low sequence identity (<40%)	protein
2	Protein Sequence Classification	High sequence indentiy (≥40%)	protein
3	Gene Tree Inference	SwissTree	protein
4	Genome-based Phylogeny	29 E.coli/Shigella strains	unassembled reads
		29 E.coli/Shigella strains	full genomes
		25 fish mitochondrial genomes	full genomes
		14 plant species	unassembled reads
		14 plant species	full genomes
5	Horizontal Gene Transfer	27 E.coil/Shigella strains	full genomes
		8 Yersinia species	full genomes
		33 simulated genomes	full artifical genomes

Viewing benchmark results

AFproject provides performance reports of alignment-free methods for every reference data set. Click on a bar to see the report.

Testing your own method

Provide us your own pairwise distances for given data set. After uploading your predictions, you will be able to compare the quality of your predictions against other alignment-free methods.

Download data set

Alignment-free method developer downloads from the AFproject server sequence set from one reasearch category.

E.g. fish mitochondiral genomes: See data set

Step 1

Calculate sequence distances

Developer uses the downloaded dataset as an input to his/her alignment-free program.

The output file should contain all-versus-all pairwise sequence distances in TSV/PHYLIP.

Step 2

Upload output

Developer uploads the output file to the server through form on the web site.

E.g. upload distances for mtDNA data set using the form

Step 3

Automatic Benchmark

The server benchmarks the uploaded predictions and presents a report with the submitted method's performance and comparison to other available methods.

By default, the report is private — visible only to the submitter — and the developer can choose if make the report publicly available or delete it.

Step 4

If you use the AFproject benchmarking service in your work, please consider citing:

Zielezinski et al. Benchmarking of alignment-free sequence comparison methods. Genome Biology, 2019, 20, p144. doi: 10.1186/s13059-019-1755-7