search

vibansal / HapCUT2

software tools for haplotype assembly from sequence data
BSD 2-Clause "Simplified" License
207 stars 36 forks source link

split blocks #117

Closed ruqianl closed 3 years ago

ruqianl commented 3 years ago

Hi,

Thanks for the great tool. I'm a bit confused from reading the methods and implementation.

is the --split_blocks option removed? With the current implementation, is there an argument that tells HAPCUT2 to split into separate blocks if the confidence is low?

Thanks, Ruqian

ruqianl commented 3 years ago

I found actually just by specifying --split_threshold and --error-analysis-mode 0 would generate splitted blocks.

ruqianl commented 3 years ago

Hi,

I found the logic of --split_blocks and --error_analysis_mode a bit off.

So I see that you have removed the argument --split_blocks and SPLIT_BLOCKS is now always set as 0.

Then there is no way of splitting the blocks anymore, we can only get the posterior probs as the scores but are not able to tell HapCUT2 to split it. Is this the intended behaviour? Sorry if I have been missing something.

Thanks, Ruqian

vibansal commented 3 years ago

Sorry for the late response. The split_blocks option is not intended for normal use and has not been tested extensively, therefore, it has been removed from the command line options. You can still use it by editing the code in the file optionparser.c. The error_analysis_mode doesn't work well and is not recommended.

ruqianl commented 3 years ago

Thanks @vibansal , I did compile a version with --split_blocks enabled, and from my observation on my datasets, it sometimes triggers an exception, but haven't had the chance to look into that yet.