pf std from interop

[katiemharding]

I know that I can get % clusters passing filter mean with:

percent_pf = summary.at(read_index).at(lane_index).at(surface_index).percent_pf().mean()

but trying to get the standard deviation of the same doesn't work percent_pf_std = summary.at(read_index).at(lane_index).at(surface_index).percent_pf().std()

Base Space gives something like: Lane 1 Read 1 Cluster PF(%) 81.26 -/+ 1.64. I am looking for how to get the 1.64 part

[ezralanglois]

First, BaseSpace is showing you per lane per read. You are accessing the data with per Lane per Read per Surface.

And what does not work mean? Are you getting NaN?

[katiemharding]

percent_pf = summary.at(read_index).at(lane_index).at(surface_index).percent_pf().mean() returns 74.40805053710938

percent_pf_std = summary.at(read_index).at(lane_index).at(surface_index).percent_pf().std() returns:

---------------------------------------------------------------------------
AttributeError                            Traceback (most recent call last)
Input In [19], in <cell line: 1>()
----> 1 test.std()

File ~/.virtualenvs/reagent-qc-xvEGZSY3-py3.9/lib/python3.9/site-packages/interop/py_interop_summary.py:409, in metric_stat.<lambda>(self, name)
    407 __setattr__ = lambda self, name, value: _swig_setattr(self, metric_stat, name, value)
    408 __swig_getmethods__ = {}
--> 409 __getattr__ = lambda self, name: _swig_getattr(self, metric_stat, name)
    410 __repr__ = _swig_repr
    412 def __init__(self, *args):

File ~/.virtualenvs/reagent-qc-xvEGZSY3-py3.9/lib/python3.9/site-packages/interop/py_interop_summary.py:80, in _swig_getattr(self, class_type, name)
     78 if method:
     79     return method(self)
---> 80 raise AttributeError("'%s' object has no attribute '%s'" % (class_type.__name__, name))

AttributeError: 'metric_stat' object has no attribute 'std'

I would like it to return some sort of a value.

[katiemharding]

for read 1 percent_pf = summary.at(read_index).at(lane_index).at(surface_index).percent_pf().mean() returns 82.2605, 80.2609, 82.2605, 80.2609,82.2605, 80.2609,82.2605, 80.2609, with a mean of 81.2607 (which matches base space) I want the standard deviation, which base space returns as 1.65, and I can't calculate from what percent_pf = summary.at(read_index).at(lane_index).at(surface_index).percent_pf().mean() returns.

[ezralanglois]

This error is saying that std is not the name of the method you need:

AttributeError: 'metric_stat' object has no attribute 'std'

You should replace it with stddev

https://github.com/Illumina/interop/blob/c98d2689941cd557e6dad43884ff12b55b3e327b/interop/model/summary/metric_stat.h#L93

[katiemharding]

Thank you!
This is great, and a lot closer to what is displayed on the basespace page. BaseSpace shows 81.26 +/- 1.64 using .percent_pf().mean() I get 81.62. but .percent_pf().stddev() I get 1.29 (it actually returns two values, 1.169 and 1.419).

Can I use this to get the basespace %PF mean and Stdev?

[ezralanglois]

This is not what BaseSpace uses

percent_pf = summary.at(read_index).at(lane_index).at(surface_index).percent_pf().mean()
percent_pf_std = summary.at(read_index).at(lane_index).at(surface_index).percent_pf().stddev()

This is what BaseSpace uses, when it prints something like this: Lane 1 Read 1 Cluster PF(%) 81.26 -/+ 1.64.

percent_pf = summary.at(read_index).at(lane_index).percent_pf().mean()
percent_pf_std = summary.at(read_index).at(lane_index).percent_pf().stddev()

Depending on the flow cell type, these can give you the same or different answers because the first reports on the surface level and the second on the lane level.

[katiemharding]

awesome! That works, and the data matches. Thank you.