Coupled autoencoders for M, E, and T analysis

Objectives:

• Joint analysis of Morphology, Electrophysiology, and Transcriptomic data from Patch-seq experiments.
• Extending results from Patch-seq dataset to EM reconstructions

Data

• Patch-seq dataset for V1 cortical interneurons (Gala et al. 2021: 3411 cells in T and E)
• Patch-seq dataset for V1 cortical interneurons (Gouwens et al. 2020: 3819 cells in T and E)
• Density representations for morphology (721 cells)

Environment

1. Navigate to the cplAE_MET folder with the setup.py file.
2. Create the conda environment with the main dependencies.

conda create -n cplAE_MET
conda activate cplAE_MET
conda install python=3.8
conda install pytorch torchvision torchaudio cudatoolkit=11.3 -c pytorch #see system specific instructions
pip install scikit-learn jupyterlab seaborn pandas rich tqdm timebudget statsmodels umap-learn
pip install tensorboard

3. Install the development version of this repository

pip install -e .

4. Install the cplAE_TE repository after cloning it.

# can do this within any directory on local machine
git clone https://github.com/AllenInstitute/coupledAE-patchseq
cd coupledAE-patchseq
pip install -e .

Experiments

"T_ME_aT_5-0_aM_5-0_asd_1-0_aE_5-0_aME_5-0_lambda_ME_T_1-0_lambda_tune_ME_T_0-75_lambda_ME_M_1-0_lambda_ME_E_1-0_aug_dec_1_Enoise_0-05_Mnoise_0-0_scale_0-3_ld_5_ne_50000_ri_0_fold_2.pkl"

Additional repositories

• celltype_hierarchy - Helpers for dendrogram manipulation
• cplAE_TE - Coupled autoencoders for T and E

Config

# config.toml contents
package_dir = '/Local/code/cplAE_MET/'
MET_inh_data = '/Local/data/inh_MET_model_input_mat.mat'

# config_preproc.toml contents
package_dir = '/Users/fahimehb/Documents/git-workspace/cplAE_MET/'

#For T
specimen_ids_file = "exc_inh_specimen_ids_30Mar22.txt"
gene_file = "good_genes_beta_score.csv"
t_data_output_file = "T_data_30Mar22.csv"
t_anno_output_file = "T_anno_30Mar22.csv"
gene_id_output_file = "gene_ids_30Mar22.csv"

#For M
m_data_folder = 'm_data'
m_anno = 'm_anno.csv'
hist2d_120x4_folder = 'hist2d_120x4'
m_output_file = 'M_data_30Mar22.mat'

#For E
E_timeseries_file = "fv_Ephys_timeseries_30Mar22.h5"
ipfx_features_file = "ipfx_features_30Mar22.csv"
e_output_file = "E_data_30Mar22.csv"

#For MET
met_output_file = "MET_data_30Mar22.mat"

Contributors

Fahimeh Baftizadeh, Rohan Gala