Command Builder — Majiq Documentation

Are you planning to quantify differences in splicing between condition(s), or (B) are you planning to compare splicing levels of sample(s) within a group(s) (or quantify a single sample) ?

Quantify differences -> deltapsi or heterogen

Compare splicing levels -> psi

No, I want something else

If A, are your experimental groups composed of heterogeneous samples?

Yes

No

If A, are your experimental groups composed of heterogeneous samples?

Yes

No

If C, Are you trying to figure out how to use voila view?

Yes

No

What kind of analysis are you performing?

Comparison between two groups or more groups (e.g. WT vs. KO, No treatment vs drug1 vs drug2)

Quantification of multiple individual samples

How do you want PSI values to be calculated?

Group level PSI where replicates within a group from the build are quantified together

Individual sample PSI where every sample from every group is quantified separately

Is your dataset heterogeneous or well controlled?

Heterogeneous

Well controlled

Only show junctions that differ between groups by at least (in PSI)

Do you want to quantify intron retention?

Yes

No

Do you want to quantify de novo (unannotated) events?

Yes

No

In what fraction of samples (in %, or count) in a group do you want a junction to be quantifiable in order to appear in the results?

Default: 50%

% of experiments

or

experiments

How many samples do you need to believe a junction/intron is real enough to include for quantification (making your model more complex)?

%

Which junctions and introns are biologically relevant for your analysis? By relevant, we mean: junctions and introns must have at least X% PSI in at least Y% (or Z number) of samples per build group.

All junctions are relevant

Junctions and introns with less than X% PSI are irrelevant

%

*Note: The default is set to be relatively conservative, specifically over 1/2 of the samples in a group should have it as quantifiable. Consider changing it if you want to capture more LSV, for example when coverage is low or when you expect only a subset of the samples to contain the LSV (e.g. disease subtype). However, keep in mind that lowering that threshold may also introduce more "noise" (e.g. irrelevant junctions/LSVs) into your downstream analysis.

Have you processed any of the input samples before with the same version of MAJIQ? (and with intron retention enabled if you want to do intron retention this time)

Yes

No

How many threads do you want to use?

I don't know

What genome assembly are you using to link VOILA output with UCSC browser coordinates (common genomes only)?

Enter full path to the reference transcriptome annotation (in GFF3 format)

Enter full path(s) to the directory containing your .bam and .bam.bai files. Place multiple directories on separate lines.

If using --incremental, enter the full path(s) to the directory containing your .sj files. Place multiple directories on separate lines.

Enter full path of directory where you want to output the splicegraph and majiq files

Enter full path to configuration file (should have content from above)

How many biological groups are you analyzing?

Enter the name of each group, followed by the list of samples for that group, separated by comma

Error: group names should only contain alphanumeric and underscore ("_") characters. You may consider using the following type of algorithm to automatically clean your group names if needed. (Python3)

                                      import re
                                      def clean_group_name(x):
                                        x = re.sub(r"[()`\[\]{}\"'<>]", "", x)
                                        x = re.sub(r"[^0-9a-zA-Z]+", "_", x)
                                        return x
                                      clean_group_name("«insert your name here»")

Would you like to visualize the final results interactively, or dump them to a text TSV file for easier parsing?

Visualize in the interface

Dump to TSV

Command Builder¶