Logical-Level Building Blocks

The Map function takes a reference file, an output filename, the chromosome strand on which the operation should be applied, and an aggregation function. See the following example, where the highlighted intervals from the samples are determined as intervals overlapping with the given reference interval.

The Map function is exposed via Di4BCLI (the command line interface to Di4's bioinformatics wrapper). The Di4BCLI defines the following syntax for the Mapfunction:

Map reference results_file strand aggregate_function

For instance:

Map C:\Users\Vahid\Desktop\Di4Data\ref.narrowpeak res.bed * count

Where * means “un-stranded” as the provided input data do not define strand. count is an aggregation function, which counts the number of intervals overlapping a reference interval. As a result of executing this command, Di4BCLI shows the following message.

> Map C:\Users\Vahid\Desktop\Di4Data\ref.narrowpeak res.bed * count
 Loaded #i:    196,180     ET: 00:00:01.0622689     Speed:   184,680 #i\sec
    Map #i:    196,180     ET: 00:00:00.2214104     Speed:   886,047 #i\sec
 Export #i:    196,111     ET: 00:00:00.1353136     Speed: 1,449,307 #i\sec
-: Done  ...       Overall ET: 00:00:01.4325864

Regarding the runtime; Di4’s runtime is the runtime of the Map function, which is 00:00:00.2214104 in this run. The load and export times are Di4B and Di4BCLI runtime for loading the reference sample, calling independent instances of Di4 instances (Chr-level degree of parallelism), and saving the results.

The Cover function finds regions on domain where a particular number of intervals overlap. See the following exmaple where the Cover function has determined regions with at least 2 and at most 5 intervals overlapping:

The Cover function is exposed via Di4BCLI, and has the following syntax:

Cover results_file strand min_acc max_acc aggregate_function

For instance:

Cover res.bed * 2 5 count

As a result of executing this command, Di4BCLI shows a message similar to the following.

> Cover res.bed * 2 5 count
  Cover #b: 119,435,300     ET: 00:00:49.4589017    Speed: 2,414,839 #b\sec
-: Done  ...        Overall ET: 00:00:49.4591127

where #b shows the number of processed snapshots (119,435,300 in this example), and query processing speed is given in snapshot per second (2,414,839 #b\sec in this example).

The Summit function is a variant of Cover function, and finds regions where interval accumulation is locally maximized. See the following exmaple where the Summit function has determined regions with at least 2 and at most 5 intervals overlapping:

The Summit function is exposed via Di4BCLI, and has a syntax similar to the Cover function as the following:

Summit results_file strand min_acc max_acc aggregate_function

The AccHis function determines a histogram of interval accumulation. See the following exmaple:

The AccHis function is exposed via Di4BCLI, and has the following syntax:

AccHis results_file