diff options
author | Muriithi Frederick Muriuki | 2021-11-20 18:06:32 +0300 |
---|---|---|
committer | GitHub | 2021-11-20 18:06:32 +0300 |
commit | 92d5766f5514181cd6aa82fc0d0f225666e892cb (patch) | |
tree | 74840e8e2118e24e1f49eb780ca0bbf24704e510 /gn3 | |
parent | abc0d36f39c691652fee81bce808d625fc368e72 (diff) | |
parent | 08c81b8892060353bb7fb15555875f03bbdcb46e (diff) | |
download | genenetwork3-92d5766f5514181cd6aa82fc0d0f225666e892cb.tar.gz |
Merge pull request #56 from genenetwork/partial-correlations
Partial correlations
Diffstat (limited to 'gn3')
-rw-r--r-- | gn3/computations/partial_correlations.py | 95 | ||||
-rw-r--r-- | gn3/data_helpers.py | 15 | ||||
-rw-r--r-- | gn3/settings.py | 2 |
3 files changed, 86 insertions, 26 deletions
diff --git a/gn3/computations/partial_correlations.py b/gn3/computations/partial_correlations.py index 07dc16d..4f45159 100644 --- a/gn3/computations/partial_correlations.py +++ b/gn3/computations/partial_correlations.py @@ -5,10 +5,14 @@ It is an attempt to migrate over the partial correlations feature from GeneNetwork1. """ +import math from functools import reduce -from typing import Any, Tuple, Sequence +from typing import Any, Tuple, Union, Sequence from scipy.stats import pearsonr, spearmanr +import pandas +import pingouin + from gn3.settings import TEXTDIR from gn3.data_helpers import parse_csv_line @@ -152,7 +156,7 @@ def tissue_correlation( "Method must be one of: {}".format(",".join(method_fns.keys()))) corr, pvalue = method_fns[method](primary_trait_values, target_trait_values) - return (round(corr, 10), round(pvalue, 10)) + return (corr, pvalue) def batch_computed_tissue_correlation( primary_trait_values: Tuple[float, ...], target_traits_dict: dict, @@ -245,7 +249,7 @@ def compute_partial_correlations_fast(# pylint: disable=[R0913, R0914] all_target_trait_names: Tuple[str, ...] = processed_trait_names_values[0] all_target_trait_values: Tuple[float, ...] = processed_trait_names_values[1] - all_correlations = determine_partials( + all_correlations = compute_partial( primary_vals, control_vals, all_target_trait_names, all_target_trait_values, method) ## Line 772 to 779 in GN1 are the cause of the weird complexity in the @@ -258,32 +262,71 @@ def compute_partial_correlations_fast(# pylint: disable=[R0913, R0914] else fetched_correlations[corr[0]][0:2]) for idx, corr in enumerate(all_correlations)) -def partial_correlation_matrix( +def build_data_frame( xdata: Tuple[float, ...], ydata: Tuple[float, ...], - zdata: Tuple[float, ...], method: str = "pearsons", - omit_nones: bool = True) -> float: + zdata: Union[ + Tuple[float, ...], + Tuple[Tuple[float, ...], ...]]) -> pandas.DataFrame: """ - Computes the partial correlation coefficient using the - 'variance-covariance matrix' method - - This is a partial migration of the - `web.webqtl.correlation.correlationFunction.determinPartialsByR` function in - GeneNetwork1, specifically the `pcor.mat` function written in the R - programming language. + Build a pandas DataFrame object from xdata, ydata and zdata """ - return 0 - -def partial_correlation_recursive( - xdata: Tuple[float, ...], ydata: Tuple[float, ...], - zdata: Tuple[float, ...], method: str = "pearsons", - omit_nones: bool = True) -> float: + x_y_df = pandas.DataFrame({"x": xdata, "y": ydata}) + if isinstance(zdata[0], float): + return x_y_df.join(pandas.DataFrame({"z": zdata})) + interm_df = x_y_df.join(pandas.DataFrame( + {"z{}".format(i): val for i, val in enumerate(zdata)})) + if interm_df.shape[1] == 3: + return interm_df.rename(columns={"z0": "z"}) + return interm_df + +def compute_partial( + primary_vals, control_vals, target_vals, target_names, + method: str) -> Tuple[ + Union[ + Tuple[str, int, float, float, float, float], None], + ...]: """ - Computes the partial correlation coefficient using the 'recursive formula' - method + Compute the partial correlations. - This is a partial migration of the - `web.webqtl.correlation.correlationFunction.determinPartialsByR` function in - GeneNetwork1, specifically the `pcor.rec` function written in the R - programming language. + This is a re-implementation of the + `web.webqtl.correlation.correlationFunction.determinePartialsByR` function + in GeneNetwork1. + + This implementation reworks the child function `compute_partial` which will + then be used in the place of `determinPartialsByR`. + + TODO: moving forward, we might need to use the multiprocessing library to + speed up the computations, in case they are found to be slow. """ - return 0 + # replace the R code with `pingouin.partial_corr` + def __compute_trait_info__(target): + primary = [ + prim for targ, prim in zip(target, primary_vals) + if targ is not None] + datafrm = build_data_frame( + primary, + [targ for targ in target if targ is not None], + [cont for i, cont in enumerate(control_vals) + if target[i] is not None]) + covariates = "z" if datafrm.shape[1] == 3 else [ + col for col in datafrm.columns if col not in ("x", "y")] + ppc = pingouin.partial_corr( + data=datafrm, x="x", y="y", covar=covariates, method=method) + pc_coeff = ppc["r"] + + zero_order_corr = pingouin.corr( + datafrm["x"], datafrm["y"], method=method) + + if math.isnan(pc_coeff): + return ( + target[1], len(primary), pc_coeff, 1, zero_order_corr["r"], + zero_order_corr["p-val"]) + return ( + target[1], len(primary), pc_coeff, + (ppc["p-val"] if not math.isnan(ppc["p-val"]) else ( + 0 if (abs(pc_coeff - 1) < 0.0000001) else 1)), + zero_order_corr["r"], zero_order_corr["p-val"]) + + return tuple( + __compute_trait_info__(target) + for target in zip(target_vals, target_names)) diff --git a/gn3/data_helpers.py b/gn3/data_helpers.py index d3f942b..b72fbc5 100644 --- a/gn3/data_helpers.py +++ b/gn3/data_helpers.py @@ -24,6 +24,21 @@ def partition_all(num: int, items: Sequence[Any]) -> Tuple[Tuple[Any, ...], ...] in reduce( __compute_start_stop__, iterations, tuple())]) +def partition_by(partition_fn, items): + """ + Given a sequence `items`, return a tuple of tuples, each of which contain + the values in `items` partitioned such that the first item in each internal + tuple, when passed to `partition_function` returns True. + + This is an approximation of Clojure's `partition-by` function. + """ + def __partitioner__(accumulator, item): + if partition_fn(item): + return accumulator + ((item,),) + return accumulator[:-1] + (accumulator[-1] + (item,),) + + return reduce(__partitioner__, items, tuple()) + def parse_csv_line( line: str, delimiter: str = ",", quoting: Optional[str] = '"') -> Tuple[str, ...]: diff --git a/gn3/settings.py b/gn3/settings.py index 0a6d03b..0ac6698 100644 --- a/gn3/settings.py +++ b/gn3/settings.py @@ -52,3 +52,5 @@ CORS_HEADERS = [ GNSHARE = os.environ.get("GNSHARE", "/gnshare/gn/") TEXTDIR = f"{GNSHARE}/web/ProbeSetFreeze_DataMatrix" + +ROUND_TO = 10 |