Bedboss: testing with bedboss

[xuebingjie1990]

I tested bedboss with the tutorial files. It works without any error but there are some problems and questions I have @Khoroshevskyi :

• [x] Need to fix: the 'median_tss_dist' column is empty
• [x] Need to fix: 'other' columns are not reported in the database (columns are empty) -- this is b/c the we are no longer calling the looper pre-submit function to generate the sample yaml file, where the bedstat read from for the other column
• [x] filter out local path in the sample yaml
• [x] the 'bed_files' and 'bigbed_files' are now renamed to 'files_bed' and 'files_bigbed'. is there a reason? if not, can we keep the way it is? other wise there are lots of other places need to change (ex s2 bucket, and all the existing bedfiles on bedbase ).
• [x] the 'bed_files' and 'bigbed_files' are now inside the output dir. is that intentional?
• [x] all the bed files and bigbed files are now named by the {sample_name} instead of their original name. is that also intentional?

10/20 update:

• [x] I don't think we want standard_genome_name() function to limited the input to hg38, hg19, and mm10
• [x] About the get_osm_path() function: the R script has tryCatch to handle missing files needed for GD plots. we don't want to kill the pipeline at the beginning because file required for one GD plot is missing

1/26 update:

• [x] do we want to add bedbuncher and bedembed as part of bedboss? @Khoroshevskyi

[khoroshevskyi]

2. I will change names of the folders so they will have previous names
3. I think, It's more convenient to sotore all the outputs, and files in one folder. And bigboss will handle it. Is it a problem?
4. No, it's not intentional. Does the names of the file should be the same ?

[xuebingjie1990]

2. I will change names of the folders so they will have previous names

okay

3. I think, It's more convenient to sotore all the outputs, and files in one folder. And bigboss will handle it. Is it a problem?

if you want to keep everything in the output dir, we will need to make changes to the metadata of all current files in the database.

4. No, it's not intentional. Does the names of the file should be the same ?

I just feel the name should be meaningful. We could use {sample_name}. Then when we make the sample table, we have to put something other than 'bedbase_demo_1', 'bedbase_demo_2', 'bedbase_demo_3'......

[xuebingjie1990]

for 1) I fixed the empty median_tss_dist. The other column is empty because you are no longer write the sample yaml file (and the log file) in the bedstat_pipeline_logs/submission dir.

[khoroshevskyi]

ok, all errors seams to be solved

[xuebingjie1990]

ok, all errors seams to be solved

okay i'll test again, with files other than hg38

also do you think we should just write the logs in the file like before? now the bedstat log is printing out on the screen. it's just very long for each sample......

[khoroshevskyi]

Printing logs is not the issue, but Yes, Wecan do it like before

[xuebingjie1990]

@khoroshevskyi can we close this issue?

[khoroshevskyi]

It is old issue, and it was solved for previous bedboss version