10X biology and sequencing model fit results

Description

We fit subsets of data provided at https://data.caltech.edu/records/2017 using a Markov modeling of mRNA transcription, splicing, degradation, and sequencing. The following files are provided: raw_results.tar.gz: raw result files output by fitting each dataset. These files should be analyzed by using the import_datasets functionality. processed_results.tar.gz: a pickle dump of ResultData structures after importing the raw results, identifying the optimal grid point, a subset of genes by the chi-squared genes, and computing Hessians for all genes at the optimal grid point. controls.tar.gz: a series of raw result files output by fitting the 10X 10k PBMC dataset with slightly different parameters. These are: --gg_210523_..._1.pickle: 1 search iteration, method of moments (MoM) initialization. --gg_210523_..._2.pickle: 1 search iteration, random initialization. --gg_210523_..._3.pickle: 20 search iterations, 1 at MoM and 19 at random parameter vectors. --gg_210524_..._4.pickle: 20 search iterations, all at random starting points. --gg_210628_..._4.pickle: Same as the first, but with a length-independent model of unspliced transcript sequencing, and identical bounds on spliced and unspliced sampling rates.

Files

Additional details