BioProfiling.jl API
Table of contents
- BioProfiling.jl API
- BioProfiling.jl examples
- Starting a codespace and setting up the environment
- Setting up Julia for package development
- Opening a pull request
- BioProfiling.jl examples
- BioProfiling.jl
Functions
Base.filter!
— MethodFor an experiment e
, update in place e.selected_features
and e.selected_entries
based on an array arr
of feature selectors and entry filters. Filters and selectors are applied sequentially.
For an experiment e
, update in place e.selected_features
and e.selected_entries
based on a feature selector s
.
For an experiment e
, update in place e.selected_features
and e.selected_entries
based on an entry filter f
.
BioProfiling.characteristic_features
— MethodReturn (all or if provided the top
) features varying the most in e
(largest absolute log fold change), when comparing entries matching filters f1
and f2
. Columns for which the fold change is negative come last.
BioProfiling.decorrelate!
— MethodReturns column of selected data in Experiment e
that are never pairwise-correlated more than 'threshold', prioritizing columns by a given order 'ordercol' (defaults to left to right).
BioProfiling.decorrelate
— MethodReturns column of 'data' that are never pairwise-correlated more than 'threshold', prioritizing columns by a given order 'ordercol' (defaults to left to right).
BioProfiling.decorrelate_by_mad!
— MethodReturns column of selected data in Experiment e
that are never pairwise-correlated more than 'threshold', prioritizing columns with largest median absolute deviation.
BioProfiling.diagnostic
— MethodReturn information in an experiment e
in column(s) feature
after filtering with filter f
.
BioProfiling.diagnostic_images
— MethodGet images in an experiment e
whose location is stored in feature s
after filtering with filter f
. Images will be saved at the path
provided if saveimages
is set to true, while keeping up to keepsubfolders
folders from the original folder structure of the images. If center
is true, crosses will indicate the center of selected objects, assuming they are stored in variables :AreaShape_Center_X
, :AreaShape_Center_X_1
, :AreaShape_Center_Y
and :AreaShape_Center_Y_1
(default if you merge nuclear and cytoplasm measurements from CellProfiler outputs). Display up to showlimit
images if show
is true. Save up to savelimit
images if saveimages
is true. If rgx
provides a list of regex substitutions, it will be applied on all image paths (which is useful if you're in a different file system or environment as the one described in the Experiment
's data). If rgb
provides a list of 3 lists of regex substitutions, it will be applied to generate the path to 3 images.
BioProfiling.diagnostic_path
— MethodGet path to images in an experiment e
stored in feature s
after filtering with filter f
. If center
is true, the path will be returned together with the center of the filtered entries, assuming they are stored in variables AreaShape_Center_X
, AreaShape_Center_X_1
, AreaShape_Center_Y
and AreaShape_Center_Y_1
. If rgx
provides a list of regex substitutions, it will be applied on all file paths in the output (which is useful if you're in a different file system or environment as the one described in the Experiment
's data).
BioProfiling.distance_mahalanobis_center
— MethodCompute the Mahalanobis Distance to center (MDC) in a dataset 'data' for a given perturbation of indices 'indpert' compared to a reference of indices 'indref'.
BioProfiling.distance_mahalanobis_median
— MethodCompute the median Mahalanobis Distance (MD) in a dataset 'data' for a given perturbation of indices 'indpert' compared to a reference of indices 'indref'.
BioProfiling.distance_robust_hellinger
— MethodCompute the Robust Hellinger Distance (RHD) in a dataset data
for a given perturbation of indices indpert
compared to a reference of indices indref
. This function calls R using RCall and a seed is set by default to ensure the results are reproducible. If you don't want that, for instance if you also use RCall and rely on another seed, set 'r_seed' to false.
BioProfiling.distance_robust_mahalanobis_median
— MethodCompute the median Robust Mahalanobis Distance (RMD) in a dataset 'data' for a given perturbation of indices 'indpert' compared to a reference of indices 'indref'. See https://e-archivo.uc3m.es/bitstream/handle/10016/24613/ws201710.pdf This function calls R using RCall and a seed is set by default to ensure the results are reproducible. If you don't want that, for instance if you also use RCall and rely on another seed, set 'r_seed' to false.
BioProfiling.filter_entries!
— MethodFilter entries in an Experiment e
based on filter(s) f
, updating e.selected_entries
in place accordingly.
BioProfiling.filter_entries
— MethodReturn filtered entries in an Experiment e
based on filter f
BioProfiling.getdata
— MethodReturn a copy of the data in Experiment e
for its selected entries and features.
BioProfiling.hellinger
— MethodSquared hellinger distance for covariance estimators S and centers µ
BioProfiling.logtransform!
— MethodApproximate normal distribution of selected entries for all selected features of an Experiment e
. Warning: columns are converted to float when necessary.
BioProfiling.logtransform
— MethodApproximate normal distribution
BioProfiling.mahalanobis
— MethodSquared mahalanobis distance for covariance estimator S and center µ
BioProfiling.most_correlated
— MethodReturn (all or if provided the top
) features in e
associated the most with ref
(absolute Pearson correlation).
BioProfiling.most_variable_features
— MethodReturn the features of e
ranked by decreasing median absolute deviation. Trim to the first top
features if a value is provided.
BioProfiling.negation
— MethodReturn a negative Filter or Selector by inverting the entries or features that are kept and excluded.
BioProfiling.normtransform!
— MethodCenter and scale all selected entries for each selectead features of an Experiment e
on control values matching a Filter f
, based on the median and median absolute deviation of the control. Warning: columns are converted to float when necessary.
BioProfiling.normtransform
— MethodCenter and scale on control values
BioProfiling.robust_morphological_perturbation_value
— FunctionCompute the Robust Morphological Perturbation Value (RMPV) for a given Experiment e
, for all levels of a column s
, compared to rows matching a given filter f
or where s
is equal to ref
. The RMPV quantifies the significance of changes between all conditions (levels in s
) and a reference condition (defined by the filter f
). In brief, the distance of type dist
between points of each perturbation and points of the reference is computed and its statistical significance is defined using a permutation test in which the perturbation and reference labels are shuffled nb_rep
times. If process_pool
is a pool of worker processes, they will be used for parallel computation in the permutation test. This function might call R using RCall and a seed is set by default to ensure the results are reproducible. If you don't want that, for instance if you also use RCall and rely on another seed, set 'r_seed' to false. This returns a DataFrame with three columns:
Condition
: the levels ins
Distance
: the distance between a condition and the
reference
RMPV
: the RMPV (empirical p-value corrected for multiple
testing)
BioProfiling.select_features!
— MethodReturn selected features in an Experiment e
based on selectors s
, updating e.selected_features
in place accordingly.
BioProfiling.select_features
— MethodReturn selected features in an Experiment e
based on selectors s
BioProfiling.shuffled_distance_mahalanobis_center
— MethodPermute labels and compute the Mahalanobis Distance to center (MDC) in a dataset 'data' for a given perturbation of indices 'indpert' compared to a reference of indices 'indref', to create an empirical distribution.
BioProfiling.shuffled_distance_mahalanobis_median
— MethodPermute labels and compute the median Mahalanobis Distance (RMD) in a dataset 'data' for a given perturbation of indices 'indpert' compared to a reference of indices 'indref', to create an empirical distribution.
BioProfiling.shuffled_distance_robust_hellinger
— MethodPermute labels and compute the Robust Hellinger Distance (RHD) in a dataset data
for a given perturbation of indices indpert
compared to a reference of indices indref
, to create an empirical distribution. This function calls R using RCall and a seed is set by default to ensure the results are reproducible. If you don't want that, for instance if you also use RCall and rely on another seed, set 'r_seed' to false.
BioProfiling.shuffled_distance_robust_mahalanobis_median
— MethodPermute labels and compute the median Robust Mahalanobis Distance (RMD) in a dataset 'data' for a given perturbation of indices 'indpert' compared to a reference of indices 'indref', to create an empirical distribution. This function calls R using RCall and a seed is set by default to ensure the results are reproducible. If you don't want that, for instance if you also use RCall and rely on another seed, set 'r_seed' to false.
DataFrames.select!
— MethodFor an experiment e
, update in place e.selected_features
and e.selected_entries
based on an array arr
of feature selectors and entry filters. Filters and selectors are applied sequentially.
For an experiment e
, update in place e.selected_features
and e.selected_entries
based on a feature selector s
.
For an experiment e
, update in place e.selected_features
and e.selected_entries
based on an entry filter f
.
FreqTables.freqtable
— MethodExpand freqtable
to support Experiment
objects. Find the frequency of the values taken by feature s
in Experiment e
.
FreqTables.freqtable
— MethodExpand freqtable
to support Experiment
objects. Find the frequency of the values taken by feature s
in Experiment e
.
Index
Base.filter!
BioProfiling.characteristic_features
BioProfiling.decorrelate
BioProfiling.decorrelate!
BioProfiling.decorrelate_by_mad!
BioProfiling.diagnostic
BioProfiling.diagnostic_images
BioProfiling.diagnostic_path
BioProfiling.distance_mahalanobis_center
BioProfiling.distance_mahalanobis_median
BioProfiling.distance_robust_hellinger
BioProfiling.distance_robust_mahalanobis_median
BioProfiling.filter_entries
BioProfiling.filter_entries!
BioProfiling.getdata
BioProfiling.hellinger
BioProfiling.logtransform
BioProfiling.logtransform!
BioProfiling.mahalanobis
BioProfiling.most_correlated
BioProfiling.most_variable_features
BioProfiling.negation
BioProfiling.normtransform
BioProfiling.normtransform!
BioProfiling.robust_morphological_perturbation_value
BioProfiling.select_features
BioProfiling.select_features!
BioProfiling.shuffled_distance_mahalanobis_center
BioProfiling.shuffled_distance_mahalanobis_median
BioProfiling.shuffled_distance_robust_hellinger
BioProfiling.shuffled_distance_robust_mahalanobis_median
DataFrames.select!
FreqTables.freqtable
FreqTables.freqtable