Single sample: Step 4: psi.py

[zhengshimao]

An error occurred when running Shiba on a single sample：

[2025-03-14 21:49:38,800] INFO Executing Step 4: psi.py...
[2025-03-14 21:49:40,306] INFO Starting PSI calculation
[2025-03-14 21:49:40,306] INFO Loading event and junction files...
[2025-03-14 21:49:41,244] INFO Processing group data...
Traceback (most recent call last):
File "/opt_shiba/Shiba/src/psi.py", line 216, in
main()
File "/opt_shiba/Shiba/src/psi.py", line 89, in main
junc_group_df = shibalib.sum_reads(params["onlypsi_group"], junc_df, group_df, group_list)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/opt_shiba/Shiba/src/lib/shibalib.py", line 192, in sum_reads
junc_df = junc_df[col]
~~~~~~~^^^^^
File "/usr/local/lib/python3.11/site-packages/pandas/core/frame.py", line 3813, in getitem
indexer = self.columns._get_indexer_strict(key, "columns")[1]
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/usr/local/lib/python3.11/site-packages/pandas/core/indexes/base.py", line 6070, in _get_indexer_strict
self._raise_if_missing(keyarr, indexer, axis_name)
File "/usr/local/lib/python3.11/site-packages/pandas/core/indexes/base.py", line 6133, in _raise_if_missing
raise KeyError(f"{not_found} not in index")
KeyError: "['Ref'] not in index"
[2025-03-14 21:49:43,300] ERROR Error executing Step 4: psi.py. Exiting..

[NaotoKubota]

Hi @zhengshimao ,

Could you please copy and paste your experiment table and config yaml file so that I can survey potential problems.
Thank you!

[zhengshimao]

OK!
original file name:
test_experiment.txt => test_experiment.tsv
test_Shiba.txt => test_Shiba.yaml

test_experiment.txt
test_Shiba.txt

[NaotoKubota]

Please try to set only_psi_group "True" in config.yaml:

only_psi_group:
  True

Shiba assumes at least two groups when performing differential splicing analysis.
For single sample analysis, you can disable the differential analysis mode by setting only_psi_group or only_psi True.

[zhengshimao]

The Step 4: psi.py is OK!

but a new error occurred in Step 6: pca.py。 Is this step applicable to single sample?

[2025-03-16 11:52:13,300] INFO Executing Step 6: pca.py...
[2025-03-16 11:52:14,208] INFO Starting PCA analysis...
[2025-03-16 11:52:14,209] INFO Loading input files...
[2025-03-16 11:52:14,386] INFO Performing PCA...
Traceback (most recent call last):
File "/opt_shiba/Shiba/src/pca.py", line 150, in
main()
File "/opt_shiba/Shiba/src/pca.py", line 132, in main
tpm_feature_df, tpm_contribution_df = mtx2pca(tpm_df, args.genes)
^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/opt_shiba/Shiba/src/pca.py", line 102, in mtx2pca
pca.fit(normalized_df)
File "/usr/local/lib/python3.11/site-packages/sklearn/base.py", line 1389, in wrapper
return fit_method(estimator, *args, **kwargs)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/usr/local/lib/python3.11/site-packages/sklearn/decomposition/_pca.py", line 442, in fit
self._fit(X)
File "/usr/local/lib/python3.11/site-packages/sklearn/decomposition/_pca.py", line 505, in _fit
X = validate_data(
^^^^^^^^^^^^^^
File "/usr/local/lib/python3.11/site-packages/sklearn/utils/validation.py", line 2944, in validate_data
out = check_array(X, input_name="X", **check_params)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/usr/local/lib/python3.11/site-packages/sklearn/utils/validation.py", line 1107, in check_array
_assert_all_finite(
File "/usr/local/lib/python3.11/site-packages/sklearn/utils/validation.py", line 120, in _assert_all_finite
_assert_all_finite_element_wise(
File "/usr/local/lib/python3.11/site-packages/sklearn/utils/validation.py", line 169, in _assert_all_finite_element_wise
raise ValueError(msg_err)
ValueError: Input X contains NaN.
PCA does not accept missing values encoded as NaN natively. For supervised learning, you might want to consider sklearn.ensemble.HistGradientBoostingClassifier and Regressor which accept missing values encoded as NaNs natively. Alternatively, it is possible to preprocess the data, for instance by using an imputer transformer in a pipeline or drop samples with missing values. See https://scikit-learn.org/stable/modules/impute.html You can find a list of all estimators that handle NaN values at the following page: https://scikit-learn.org/stable/modules/impute.html#estimators-that-handle-nan-values
[2025-03-16 11:52:15,497] ERROR Error executing Step 6: pca.py. Exiting...

[NaotoKubota]

This is expected because PCA needs at least two samples to run. I will fix it so that pca.py is skipped if only one sample is provided. Thanks for reporting this issue!

[zhengshimao]

Thank you for your patient answer！