--- size_categories: - 1M>> import datasets ``` The user can then load the `CF-MS_Homo_sapiens_PPI` dataset using `datasets.load_dataset(...)`. There are two configurations, or 'views' for the set. The user can choose between them via the `name` parameter: * `pairs` (Default): Pairwise protein elution profiles with binary labels for whether the two proteins are known to interact ``` >>> view = "pairs" >>> dataset = datasets.load_dataset( path = "viridono/CF-MS_Homo_sapiens_PPI", name = view) ``` * `proteins`: Individual protein elution profiles without labels for if the user wishes to assemble in a non-pairwise fashion ``` >>> view = "proteins" >>> dataset = datasets.load_dataset( path = "viridono/CF-MS_Homo_sapiens_PPI", name = view) ``` and the dataset will be loaded as a `datasets.DatasetDict`. For pairs: ``` >>> dataset DatasetDict({ train: Dataset({ features: ['experiment_id', 'uniprot_id1', 'uniprot_id2', 'elut_trace1', 'elut_trace2', 'label'], num_rows: 2496144 }) test: Dataset({ features: ['experiment_id', 'uniprot_id1', 'uniprot_id2', 'elut_trace1', 'elut_trace2', 'label'], num_rows: 2769931 }) }) ``` and for proteins: ``` DatasetDict({ train: Dataset({ features: ['experiment_id', 'uniprot_id', 'fraction_names', 'trace'], num_rows: 20383 }) }) ``` This is a column-wise format. Elution traces are 1D vectors of protein abundances (PSMs) that are stored either in the `elut_trace#` column (for pairs) or in the `trace` (for individual proteins). Note that the traces have been uploaded in a lossless format, meaning they are not normalized across different experiments (`experiment_id`) (i.e. have differing lengths, differing peak heights). The user may wish to normalize elution data when training. This is easily achievable following conversion to a `pandas.DataFrame`. Note that the `DatasetDict` must first be partitioned into its train and test splits: ``` >>> ds_train = dataset['train'] >>> ds_test = dataset['test'] >>> ds_train.to_pandas() >>> ds_test.to_pandas() ``` # Useful Pandas Normalizations / Transformations As a `pandas.DataFrame`, the user can then apply any of various transformations, including padding the 1D vectors to make them of uniform length: ``` max_len = max(df_train['elut_trace1'].apply(len)) df_train['elut_trace1'] = df_train['elut_trace1'].apply(lambda x: np.pad(x, (0, max_len - len(x)), mode='constant')) ``` or value-wise normalization, for example row-max: ``` df_train['elut_trace1'] = df_train['elut_trace1'].apply(lambda x: x / x.max() if x.max() != 0 else x) ``` Also note that elution data can be rather sparse, so the user might want to extract only the elution vectors that reach a certain minimum PSM threshold. **This should be done prior to value normalization**. Good values for minimum peak height are 5 or 10: ``` df_filtered = df_train[df_train['elut_trace1'].apply(lambda x: np.any(x >= 10)) & df_train['elut_trace2'].apply(lambda x: np.any(x >= 10))] ``` # CF/MS Elution Profile PPI Dataset Proteins are the functional basis of life, but it is often their interactions with other proteins which gives rise to said functions. Therefore, we are often interested in whether two proteins participate in the same *protein complex*, or if they **'co-complex'**. Co-fractionation mass spectrometry (CF/MS) is a high-throughput method for determining whether proteins form complexes. If they do, both proteins will typically separate out into the same fractions, or **'co-elute'**, during column chromatography experiments. As a result, their abundances will be *highly correlated* across all the fractions measured. CF/MS leverages this fact to identify new protein complexes by attempting to statistically correlate the elution profiles of groups of proteins. Typically, we use Pearson correlation coefficient to determine correlation between protein pairs. While this often works quite well, Pearson is a linear function. Current research is exploring whether there are non-linear, higher-order signals between these elution profiles that might have better predictive power than Pearson. As deep learning models excel at estimating non-linear relationships in data, the goal of this dataset is to act as training data for such models, especially **Siamese networks**. This dataset includes processed data from several *Homo sapiens* protein **co-fractionation mass spectrometry (CF/MS)** experiments, as well as positive/negative protein-protein interaction (PPI) labels for each pair. Collated, maintained by **Drew Lab at University of Illinois at Chicago** [Drew Lab webpage](https://ksdrew.github.io/) ## File formats - **The .elut file**: A **.elut** file is a TSV-like format containing raw count data from a chromatographic fractionation experiment. Each row in a **.elut** file shows the abundance of a single protein across the collected fractions (columns). Generally speaking, these fractions are collected over time. However, different chromatographic columns can separate proteins along different axes. For example, *Size-eclusion chromatography (SEC)* will mostly separate proteins into fractions according to their *size*; *Ion-exchange chromatography (IEX)* will separate them into fractions according to their *charge*. Each file in this dataset comes from one of these two column separation methods and is named accordingly ('...xx_SEC_xx...' / '...xx_IEX_xx...'). We refer to a given protein's (row's) count data across all fractions (columns) as that protein's elution trace or elution profile. To summarize: - A given row contains count data for a specific protein - A row's *first column* contains its associated **protein ID** - A row's *subsequent columns* contain that protein's count data from the fractionation experiment - **Note: The user may notice that the first row in a .elut file is one column longer than subsequent rows. This is because the first row contains row names (protein IDs), and the first column contains column names (fraction IDs). Therefore, cell 'A0' is empty.** ## File structure - The **.elut** files each contain a collection elution traces for proteins from a given CF/MS experiment. These can be paired to make sample data. A complete list of data sources can be found at the bottom of this README - The **.txt** files contain line-wise specification of protein complexes used to generate positive/negative labels. These can be used to direct the pairing of elution traces into data points. - **intact_complex_merge_20230309.train_ppis.txt**: List of positive PPIs for training data - **intact_complex_merge_20230309.test_ppis.txt**: List of positive PPIs for testing data - **intact_complex_merge_20230309.neg_train_ppis.txt**: List of negative PPIs for training data - **intact_complex_merge_20230309.neg_test_ppis.txt**: List of negative PPIs for testing data - **intact_complex_merge_20230309.train.txt** Line-wise list of protein complexes ### List of publications/experiments from which this dataset was assembled: Connelly, K. E., Hedrick, V., Paschoal Sobreira, T. J., Dykhuizen, E. C., & Aryal, U. K. (2018). Analysis of Human Nuclear Protein Complexes by Quantitative Mass Spectrometry Profiling. Proteomics, 18(11), e1700427. https://doi.org/10.1002/pmic.201700427 - T98G_glioblastoma_multiforme_cells_SEC_Conelly_2018_Bio1.elut - T98G_glioblastoma_multiforme_cells_SEC_Conelly_2018_Bio2.elut Kirkwood, K. J., Ahmad, Y., Larance, M., & Lamond, A. I. (2013). Characterization of native protein complexes and protein isoform variation using size-fractionation-based quantitative proteomics. Molecular & cellular proteomics : MCP, 12(12), 3851–3873. https://doi.org/10.1074/mcp.M113.032367 - U2OS_cells_SEC_Kirkwood_2013_rep1.elut - U2OS_cells_SEC_Kirkwood_2013_rep2.elut - U2OS_cells_SEC_Kirkwood_2013_rep3.elut Larance, M., Kirkwood, K. J., Tinti, M., Brenes Murillo, A., Ferguson, M. A., & Lamond, A. I. (2016). Global Membrane Protein Interactome Analysis using In vivo Crosslinking and Mass Spectrometry-based Protein Correlation Profiling. Molecular & cellular proteomics : MCP, 15(7), 2476–2490. https://doi.org/10.1074/mcp.O115.055467 - U2OS_cells_SEC_Larance_2016_PT3281S1.elut - U2OS_cells_SEC_Larance_2016_PT3441S1.elut - U2OS_cells_SEC_Larance_2016_PT3442S1.elut - U2OS_cells_SEC_Larance_2016_PT3701S1.elut - U2OS_cells_SEC_Larance_2016_PTSS3801.elut - U2OS_cells_SEC_Larance_2016_PTSS3802.elut Mallam, A. L., Sae-Lee, W., Schaub, J. M., Tu, F., Battenhouse, A., Jang, Y. J., Kim, J., Wallingford, J. B., Finkelstein, I. J., Marcotte, E. M., & Drew, K. (2019). Systematic Discovery of Endogenous Human Ribonucleoprotein Complexes. Cell reports, 29(5), 1351–1368.e5. https://doi.org/10.1016/j.celrep.2019.09.060 - HEK_293_T_cells_SEC_Mallam_2019_C1.elut - HEK_293_T_cells_SEC_Mallam_2019_C2.elut Moutaoufik, M. T., Malty, R., Amin, S., Zhang, Q., Phanse, S., Gagarinova, A., Zilocchi, M., Hoell, L., Minic, Z., Gagarinova, M., Aoki, H., Stockwell, J., Jessulat, M., Goebels, F., Broderick, K., Scott, N. E., Vlasblom, J., Musso, G., Prasad, B., Lamantea, E., … Babu, M. (2019). Rewiring of the Human Mitochondrial Interactome during Neuronal Reprogramming Reveals Regulators of the Respirasome and Neurogenesis. iScience, 19, 1114–1132. https://doi.org/10.1016/j.isci.2019.08.057 - NTera2_embryonal_carcinoma_stem_cells_IEX_Moutaoufik_2019_2_R1.elut - NTera2_embryonal_carcinoma_stem_cells_IEX_Moutaoufik_2019_2_R2.elut - NTera2_embryonal_carcinoma_stem_cells_IEX_Moutaoufik_2019_R1.elut - NTera2_embryonal_carcinoma_stem_cells_IEX_Moutaoufik_2019_R2.elut - NTera2_embryonal_carcinoma_stem_cells_SEC_Moutaoufik_2019_2_R1.elut - NTera2_embryonal_carcinoma_stem_cells_SEC_Moutaoufik_2019_2_R2.elut - NTera2_embryonal_carcinoma_stem_cells_SEC_Moutaoufik_2019_R1.elut - NTera2_embryonal_carcinoma_stem_cells_SEC_Moutaoufik_2019_R2.elut Wan, C., Borgeson, B., Phanse, S., Tu, F., Drew, K., Clark, G., Xiong, X., Kagan, O., Kwan, J., Bezginov, A., Chessman, K., Pal, S., Cromar, G., Papoulas, O., Ni, Z., Boutz, D. R., Stoilova, S., Havugimana, P. C., Guo, X., Malty, R. H., … Emili, A. (2015). Panorama of ancient metazoan macromolecular complexes. Nature, 525(7569), 339–344. https://doi.org/10.1038/nature14877 - CB660_neural_stem_cell_IEX_Wan_2015.elut - G166_glioma_stem_cell_IEX_Wan_2015_Hs_HCW_2.elut - G166_glioma_stem_cell_IEX_Wan_2015_Hs_HCW_3.elut - IEX_Wan_2015_Hs_HCW_4.elut - IEX_Wan_2015_Hs_HCW_5.elut - IEX_Wan_2015_Hs_HCW_6.elut - IEX_Wan_2015_Hs_HCW_7.elut - IEX_Wan_2015_Hs_HCW_8.elut - IEX_Wan_2015_Hs_HCW_9.elut - IEX_Wan_2015_Hs_IEX_1.elut - IEX_Wan_2015_Hs_IEX_2.elut