Notebooks for pre-processing, analysis, and visualization of PRESCIENT applied to in vitro hematopoeisis (Weinreb et al. 2020) and in vitro beta-cell differentiation (Veres et al. 2019).
- Current paper version: https://www.biorxiv.org/content/10.1101/2020.08.26.269332v1
- For in development PRESCIENT source code please visit https://github.com/gifford-lab/prescient/.
/scripts contains example bash scripts for fitting models:
weinreb-interpolate.shtrains 5 seeds of 2 layer 400 unit models on only lineage tracing data with ground truth proliferation rates for the interpolation task on the Weinreb et al. datasetweinreb-fate.shtrains 5 seeds of 2 layer 400 unit models on all data with estimated proliferation rates for clonal fate bias prediction on the Weinreb et al. datasetveres-fate.shtrains 5 seeds of 2 layer 400 unit models with estimated proliferation rates on the Veres et al. dataset
/notebooks contains preprocessing, analysis and visualization workflows:
02{a-e}-weinreb2020-interpolation-*contains workflows for the interpolation task on the Weinreb et al. dataset03{a-d}-weinreb2020-fate-*contains workflows for fate prediction task on the Weinreb et al. dataset04{a-d}-weinreb2020-perturbations-*contains visualizations for perturbational experiments on the Weinreb et al. dataset05{a-d}-veres2019-*contains workflows for fate prediction on the Veres et al. dataset05{c-e}-veres2019-perturbations-*contains visualizations for timing perturbations on the Veres et al. dataset.06{a-e}-veres2019-perturbations-*contains visualizations for large-scale perturbational screens and cell subset perturbations on the Veres et al. dataset.
Pre-processed data as generated by the workflows above can be downloaded from here
Pre-trained models used for perturbation experiments can be downloaded from here
React
Typescript
Django
Laravel
Facebook
Microsoft
Google
Alibaba
D3
Tencent