Running build then test then viz

[jjyotikataria]

Hi Andre

From the sample data of test case events, I took four bam files 1,2,3,4
I ran these commands:

/home/nitin.m/SPLAADER/spladder-master/Python-3.6.4/bin/spladder build -o . -a /home/jyoti.k/TESTING/SPLADDER/SPLADDER_RESULTS/EXAMPLE/spladder-master/tests/testcase_events/data/testcase_events_spladder.gtf -b bam.txt --extract-ase --event-types exon_skip,intron_retention,alt_3prime,alt_5prime,mutex_exons,mult_exon_skip --readlen 15 --output-conf-icgc --output-txt --output-txt-conf --output-gff3 --output-struc --output-struc-conf --output-bed --output-conf-bed --output-conf-tcga --parallel 30 --merge-strat merge_graphs

test command (2,2 bam files in each bam1 and bam2 text file). I had to run test command separately for each event because it was throwing error otherwise. did I do right way? Please let me know.

/home/nitin.m/SPLAADER/spladder-master/Python-3.6.4/bin/spladder test -o . -v --diagnose-plots -f pdf --merge-strat merge_graphs --readlen 50 --event-types exon_skip -a bam1.txt -b bam2.txt

/home/nitin.m/SPLAADER/spladder-master/Python-3.6.4/bin/spladder test -o . -v --diagnose-plots -f pdf --merge-strat merge_graphs --readlen 50 --event-types intron_retention -a bam1.txt -b bam2.txt
/home/nitin.m/SPLAADER/spladder-master/Python-3.6.4/bin/spladder test -o . -v --diagnose-plots -f pdf --merge-strat merge_graphs --readlen 50 --event-types alt_3prime -a bam1.txt -b bam2.txt
/home/nitin.m/SPLAADER/spladder-master/Python-3.6.4/bin/spladder test -o . -v --diagnose-plots -f pdf --merge-strat merge_graphs --readlen 50 --event-types alt_5prime -a bam1.txt -b bam2.txt
/home/nitin.m/SPLAADER/spladder-master/Python-3.6.4/bin/spladder test -o . -v --diagnose-plots -f pdf --merge-strat merge_graphs --readlen 50 --event-types mult_exon_skip -a bam1.txt -b bam2.txt
/home/nitin.m/SPLAADER/spladder-master/Python-3.6.4/bin/spladder test -o . -v --diagnose-plots -f pdf --merge-strat merge_graphs --readlen 50 --event-types mutex_exons -a bam1.txt -b bam2.txt

viz command

Can you tell me which one is the input for viz command? There are multiple events in the file but graph came for just 1 exon i.e. first event id. How to plot all?

Output files of test:
test_results_C3_exon_skip.gene_unique.tsv
test_results_C3_exon_skip.tsv
test_results_extended_C3_exon_skip.tsv
test_setup_C3_exon_skip.pickle

[akahles]

Dear @jjyotikataria,
sorry for the late reply. If still relevant, can you please share your command with me that you use for plotting? Usually, the viz mode will automatically select all relevant files from the SplAdder output directory and you can specify the event(s) of interest via the event id(s). So I would need to understand whether you are interested in plotting the test results or some other information about a specific event.

[padilr1]

@akahles Hi Andre, I'm trying to run the 'spladder test' on your sample data and I keep running into an error (see pic below). The samples doesn't seem to contain any other splicing events besides exon skipping. However, when I run 'spladder test' specifically for the 'exon_skip' event, I run into the error below. Any ideas to the potential problem?

I am running:
spladder build -a TAIR10_GFF3_genes.tiny.gff -b bam.txt -o output --merge-strat merge_graphs --output-txt --output-txt-conf --output-gff3 --output-conf-icgc --output-struc --output-struc-conf --output-bed --output-conf-bed --output-conf-tcga --parallel 8

spladder test -o output -v --diagnose-plots -f pdf --readlen 50 --event-types exon_skip -a bam1.txt -b bam2.txt

bam.txt just contains the .bam files from your sample directory, with bam1.txt and bam2.txt splitting the 4 samples into their respective conditions.

The error is:
"ValueError: zero-size array to reduction operation maximum which has no identity"

Thanks!

[akahles]

Dear @padilr1 ,

Unfortunately, testing the differential usage of a single event will not work in the current setting. SplAdder is estimating the mean-variance relationship from many different events. If only a single event is present, this will not be possible.
We are currently working on improving the testing setup, to also better handle such corner cases.

Best,
Andre