Update batch processing length normalization to match non-batch processing length normalization #441
Conversation
update help print statements
src/main.cpp
Outdated
| //<< " --error_rate Estimated error rate of long reads (required for --long)" << endl | ||
| << " --threshold Threshold for rate of unmapped kmers per read" << endl |
There was a problem hiding this comment.
Could you clarify what looks accidental?
There was a problem hiding this comment.
According to the diff on github, you are removing whitespace here:
In line 2073 you fix the alignment of 'Treat' but in line 2075, 'Threshold' is shifted to left too much.
| @@ -2180,7 +2180,7 @@ void usageTCCQuant(bool valid_input = true) { | |||
| << " (default: equivalence classes are taken from the index)" << endl | |||
| << "-f, --fragment-file=FILE File containing fragment length distribution" << endl | |||
| << " (default: effective length normalization is not performed)" << endl | |||
| << "--long Use version of EM for long reads " << endl | |||
There was a problem hiding this comment.
Could you clarify your question?
| @@ -2380,7 +2380,7 @@ int main(int argc, char *argv[]) { | |||
| if (fld_lr_c[i] > 0.5) { | |||
| //Good results with comment below. | |||
| //flensout_f << std::fabs((double)fld_lr[i] / (double)fld_lr_c[i] - index.k);//index.target_lens_[i] - (double)fld_lr[i] / (double)fld_lr_c[i] - k); // take mean of recorded uniquely aligning read lengths | |||
| flensout_f << std::fabs(index.target_lens_[i] - ((double)fld_lr[i] / (double)fld_lr_c[i]) - index.k); | |||
There was a problem hiding this comment.
Care to elaborate a bit? Ideally in comment, else maybe at least in the commit message?
There was a problem hiding this comment.
Hi! Based on our analysis for effective length normalization for long reads, the updated effective length provides better results.
There was a problem hiding this comment.
I still don't follow what you are changing and why but I am not familiar with the code. So if this makes sense to others without further explanation, feel free to ignore my comment.
| @@ -2380,7 +2380,7 @@ int main(int argc, char *argv[]) { | |||
| if (fld_lr_c[i] > 0.5) { | |||
| //Good results with comment below. | |||
| //flensout_f << std::fabs((double)fld_lr[i] / (double)fld_lr_c[i] - index.k);//index.target_lens_[i] - (double)fld_lr[i] / (double)fld_lr_c[i] - k); // take mean of recorded uniquely aligning read lengths | |||
| flensout_f << std::fabs(index.target_lens_[i] - ((double)fld_lr[i] / (double)fld_lr_c[i]) - index.k); | |||
There was a problem hiding this comment.
I still don't follow what you are changing and why but I am not familiar with the code. So if this makes sense to others without further explanation, feel free to ignore my comment.
| @@ -2180,7 +2180,7 @@ void usageTCCQuant(bool valid_input = true) { | |||
| << " (default: equivalence classes are taken from the index)" << endl | |||
| << "-f, --fragment-file=FILE File containing fragment length distribution" << endl | |||
| << " (default: effective length normalization is not performed)" << endl | |||
| << "--long Use version of EM for long reads " << endl | |||
src/main.cpp
Outdated
| //<< " --error_rate Estimated error rate of long reads (required for --long)" << endl | ||
| << " --threshold Threshold for rate of unmapped kmers per read" << endl |
There was a problem hiding this comment.
According to the diff on github, you are removing whitespace here:
In line 2073 you fix the alignment of 'Treat' but in line 2075, 'Threshold' is shifted to left too much.
No description provided.