What's new on 1.0 last release?

[seoanezonjic]

I'm user of Haploclique a long time ago and I see that you have released a major version. I didn't found a changelog but I would like to known the new features of this version. Please, can you do a brief description? Thank you in advance

[MaryamZaheri]

Please let me know if you were using the master branch, because the difference is significant. Otherwise if you were using the development branch, the new version only fixes a few bugs and adds test cases.

[seoanezonjic]

Until the last month I have been using the Armin Topfer version repo. Now, writing a paper I check haploclique on github and I saw the cbg-ethz version (master branch). I would like to know what are the new features and fixed bugs in comparison with the Topfer version. By the way, I have certain problems. First, short flags not works for me and last, I have no way to set the number of threads on haploclique and it works very slowly with one thread. Thank you in advance

[MaryamZaheri]

Main differences:

• Haploclique runs in several iterations until constructing the global super-reads or haplotypes, unlike the previous master version that required user to run iterations manually.
• After the first iteration, the super-reads are not realigned against reference genome. Instead, the sequences of alignments contributing to super-reads are directly merged. Unlike the previous version that in each iteration the super-reads were realigned to the reference genome.
• Overlapping paired-end reads were not taken into account in the original implementation of Haploclique, which are now considered.
• "max_cliques" parameter is added in order to constrain the number of cliques which are taken into the next iteration in order to reduce the running time.

There were no major bug fixing. The bugs fixed were not related to the algorithm but to validate input file. Furthermore, the current implementation is single thread.

[seoanezonjic]

Thank you by your detailed answer, Maryam. With the implementation changes that you describe I guess that time executions must be lower than with the previous version. But I have the following execution: haploclique --log=log_file --allel_frequencies=freq_file reads.bam The bam file is about 5MB (you can download it at : https://www.dropbox.com/s/rja3azyxqwvmy2v/reads.bam?dl=0 , I can't attach to this message) but the execution don't finish (it takes me days, literally). I don`t now where it is the problem. Maybe I must set the max_cliques parameter, so which value do you recommend? Also, which it is the biggest execution that you have done with haploclique (amount of reads or bam size) and how many time take the execution? Thanks in advance

[MaryamZaheri]

The default max_clique is 5000, you can also try less numbers. If you try short sequences (>2000 nc) with 200 to 500 coverage maybe you get the result faster. However, the execution time is high and we are optimizing the code to reduce it. Can you please share the allel_frequencies file also?

[seoanezonjic]

Sorry, I guessed that the allel_frequencies file was an optional output file not a input file. Maybe, this option freezes my haploclique executions? I'll try again without set this option and I'll post here the results.

[seoanezonjic]

I have performed several tests using your latest commit and the problem persists. The execution don't finish. Any recomendation to get results with this version of Haploclique? The command to execute haploclique was:

haploclique reads.bam

[MaryamZaheri]

Hi Pedro,

It is strange that you don't get result even with the default parameter. Can you send me ( maryam.zaheri@bsse.ethz.ch ) a sample of your data (reads.bam)?