training - MiraldiLab/maxATAC GitHub Wiki

Training

For this walkthrough, we will use data from the GATA3 TF model.

For the GATA3 model, we have 4 total cell types available:

  • Jurkat
  • MCF-7
  • A549
  • SK-N-SH

Using the GATA3 model, we will be able to train on 3 of the cell types and evaluate the model performance in an independently held out cell type. The data for the GATA3 model comes from multiple sources, like most maxATAC models, the data is derived from both the GEO and ENCODE database.

The following table has the metadata for experiments used:

ENCODE_ACCESSION CELL_TYPE ENCODE_BIO_REP TECH_REP BIO_REP SOURCE
NaN Jurkat 1 SRR8820028 SRX5608483 GEO
NaN Jurkat 1 SRR8820029 SRX5608483 GEO
NaN Jurkat 1 SRR8820030 SRX5608483 GEO
NaN Jurkat 1 SRR13199795 SRX9633376 GEO
NaN Jurkat 1 SRR13199794 SRX9633376 GEO
NaN Jurkat 1 SRR13199792 SRX9633375 GEO
NaN Jurkat 1 SRR13199793 SRX9633375 GEO
ENCSR422SUG MCF-7 1 SRR14103347 SRX10475183 ENCODE
ENCSR422SUG MCF-7 2 SRR14103348 SRX10475184 ENCODE
NaN MCF-7 1 SRR13199806 SRX9633382 GEO
NaN MCF-7 1 SRR13199807 SRX9633382 GEO
NaN MCF-7 1 SRR13199804 SRX9633381 GEO
NaN MCF-7 1 SRR13199805 SRX9633381 GEO
ENCSR032RGS A549 1 SRR14103424 SRX10475232 ENCODE
ENCSR032RGS A549 1 SRR14103425 SRX10475232 ENCODE
ENCSR032RGS A549 1 SRR14103426 SRX10475232 ENCODE
ENCSR032RGS A549 1 SRR14103427 SRX10475232 ENCODE
ENCSR032RGS A549 2 SRR14103420 SRX10475231 ENCODE
ENCSR032RGS A549 2 SRR14103421 SRX10475231 ENCODE
ENCSR032RGS A549 2 SRR14103422 SRX10475231 ENCODE
ENCSR032RGS A549 2 SRR14103423 SRX10475231 ENCODE
ENCSR032RGS A549 3 SRR14103428 SRX10475233 ENCODE
ENCSR032RGS A549 3 SRR14103429 SRX10475233 ENCODE
ENCSR032RGS A549 3 SRR14103430 SRX10475233 ENCODE
ENCSR032RGS A549 3 SRR14103431 SRX10475233 ENCODE
ENCSR587TRP SK-N-SH 1 SRR14305486 SRX10661032 ENCODE
ENCSR587TRP SK-N-SH 1 SRR14305487 SRX10661032 ENCODE
ENCSR587TRP SK-N-SH 1 SRR14305488 SRX10661032 ENCODE
ENCSR587TRP SK-N-SH 1 SRR14305489 SRX10661032 ENCODE
ENCSR587TRP SK-N-SH 2 SRR14305490 SRX10661033 ENCODE
ENCSR587TRP SK-N-SH 2 SRR14305491 SRX10661033 ENCODE
ENCSR587TRP SK-N-SH 2 SRR14305492 SRX10661033 ENCODE
ENCSR587TRP SK-N-SH 2 SRR14305493 SRX10661033 ENCODE
ENCSR587TRP SK-N-SH 3 SRR14305494 SRX10661034 ENCODE
ENCSR587TRP SK-N-SH 3 SRR14305495 SRX10661034 ENCODE
ENCSR587TRP SK-N-SH 3 SRR14305496 SRX10661034 ENCODE
ENCSR587TRP SK-N-SH 3 SRR14305497 SRX10661034 ENCODE