vast-tools diff failed and vast-tools compare_expr issue

[flolebill]

Dear Manuel Irimia and vast-tools team,

I am encountering other problems using vast-tools that I was not able to solve while reading previous issues on this website.

After having a problem with the combine tools, that I was able to solve thanks to Manuel, I have now a problem with the diff tool. I have performed a set of alignment, merging and combining with the will to consider both AS and expression events. When came the time to perform differential analysis, the terminal answered me that multiple warnings happened. Please see the screenshot below. I was not able to use "warning()" to see the first 50 of them.

I have then tried to use the compare_expr tool. I had an error message indicating that "vast compare needs a cRPKM + COUNTs table". I did indicate the appropriate table on the command line, as you can see below.

And the table itself is located on my "vast-out" directory, as you can see here:

I've tried to run the compare_expr tool with the "--install_limma" option, unsuccessfully.

I can now provide you some screenshots to allow you to evaluate if the files resulting from the alignement, merging and combining are good. At least, they exist and have a size indicated.

I hope that with these information, you will be able to understand my problem. I will obviously provide any supplemental information that you would need.

Thank you in advance for your help, Best regards,

Florian

[UBrau]

Hi Florian, The issue with diff looks like diff is not being started correctly. It could be an installation issue. If you just type vast-tools diff -h do you get the help text?

Ulrich

[mirimia]

Hello,

Regarding compare_expr, you're providing the cRPKM (only) table, not the cRPKM_AND_COUNTS table. To generate this, you need to use the -C option in combine. You can get it quickly (without doing the slow splicing part of combine) doing:

vast-tools combine --exprONLY -C -o ./FOLDER -sp Species

Cheers Manu

[flolebill]

Hi Florian, The issue with diff looks like diff is not being started correctly. It could be an installation issue. If you just type vast-tools diff -h do you get the help text?

Ulrich

Hi Ulrich, I do get the help text once I type vast-tools diff -h, as you can see on the following screenshots.

Florian

[flolebill]

Hello,

Regarding compare_expr, you're providing the cRPKM (only) table, not the cRPKM_AND_COUNTS table. To generate this, you need to use the -C option in combine. You can get it quickly (without doing the slow splicing part of combine) doing:

vast-tools combine --exprONLY -C -o ./FOLDER -sp Species

Cheers Manu

Hi Manuel,

Your solution did work very well. I'm surprised that there are only 4 GE events that differ in these samples but at least I can run the tool now. Thank you very much.

Florian

[mirimia]

Great. Keep in mind, though, that vast-tools is not really designed for fine GE analysis. It's just meant to give you a sense of the potential changes. (However, if you only see 4 changes here, in my experience, it's unlikely you'll find many strong changes with other software either).

[sk1350]

Hello vast-tools team, Has the vast-tools diff issue described in this post been resolved? I see the same error message: Trying to load required package: optparse Trying to load required package: MASS Trying to load required package: RColorBrewer Trying to load required package: reshape2 Trying to load required package: ggplot2 Trying to load required package: grid Trying to load required package: parallel There were 50 or more warnings (use warnings() to see the first 50)

[mirimia]

Hello,

These might be just "harmless" warnings from the R analyses. Unfortunately, diff is beyond my knowledge, but perhaps @UBrau can confirm.

Thanks Manu

[UBrau]

Hello sk1350,

Are you experiencing any actual errors (i.e., does diff not produce the expected output)? If so, please provide sample input. What you are showing looks harmless.

Ulrich

On 2/3/21 07:25, Manuel Irimia wrote:

EXTERNAL EMAIL:

Hello,

These might be just "harmless" warnings from the R analyses. Unfortunately, diff is beyond my knowledge, but perhaps @UBrau https://github.com/UBrau can confirm.

Thanks Manu

— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub https://github.com/vastgroup/vast-tools/issues/92#issuecomment-788871722, or unsubscribe https://github.com/notifications/unsubscribe-auth/AB4HSWXUGWL4YEYMQY7URVDTBTKLVANCNFSM4QEOZ66Q.

[sk1350]

Hello,

It did turn out to be a harmless error as diff did provide output. However I now have a different question. I ran vast-tools compare and it shows 0 intron retention events even though all the other AS tools I've used showed IR for this dataset. Could there have been a mistake during align or combine?

perl vast-tools compare vast_out/INCLUSION_LEVELS_FULL-xenTro3-5.tab -a stage10_1,stage10_1_1 -b stage14_1,stage14_1_1,stage15_1 --min_dPSI 10 [vast compare]: VAST-TOOLS v2.5.1 [vast compare]: Doing comparisons of AS profiles (stage10 vs stage14) [vast compare]: Printing summary statistics

Options: dPSI=10, range_dif=5, min_ALT_use=25 Summary statistics: AS_TYPE Higher_in_stage10 Higher_in_stage14 TOTAL_EV TOTAL_AS(10<PSI<90) Microexons 11 4 632 42 Long_AltEx 108 44 28212 514 Intron_ret 0 0 0 0 Alt_3ss NA 0 0 0 Alt_5ss NA 0 0 0

[mirimia]

Hello,

Yes, it looks as if you have run combine with the --onlyEX option, since there are also no Alt3/5 events. If not, maybe there has been several problems with align (but it seems less likely). The latest changes (still under v2.5.1) include some checks for incorrect align runs.

(Judging from the number of total exons, though, the coverage of the RNA-seq is not very high, so you may want to consider merging the replicates).

M

[sk1350]

Hello,

Thank you for the merging the replicates suggestion. This is an old dataset and the coverage is quite low. I didn't see anything about the --onlyEX option in the combine part of documentation, is it standard or has to be specified?

Sofia

[mirimia]

Hi Sofia,

It has to be specified, so it's strange then (it's explain in the help message of vast-tools combine). I recommend you pull the latest version of vast-tools and then run combine and compare again, just in case.

If you want, you can also do a ls -l to the to_combine/ folder, paste it here, and I'll see if the files look OK.

M

[mirimia]

Hi again,

Have you checked the information provided by vast-tools combine (in the standard output) when you ran it? It should have probably informed you about possible errors (i.e. that some tables were missing).

M

[sk1350]

Hello,

Thank you for your suggestions. So I tried combining my files and checking I do not use the --onlyEX option but when I run vast-tools combine -sp Xt1 -o /mnt/lustre/users/sk1350/tropicalis/vast-tools/vast_out -IR_version 2 -extra_eej 5

I get this error:

[vast combine error]: in call /mnt/lustre/users/sk1350/tropicalis/vast-tools/bin/RI_MakeTablePIR.R --verbose 1 -s /mnt/lustre/users/sk1350/tropicalis/vast-tools/VASTDB/Xt1 --IR_version 2 -c /mnt/lustre/users/sk1350/tropicalis/vast-tools/vast_out/to_combine -q /mnt/lustre/users/sk1350/tropicalis/vast-tools/vast_out/to_combine/Coverage_key_v2-Xt17.IRQ -o /mnt/lustre/users/sk1350/tropicalis/vast-tools/vast_out/raw_incl /usr/bin/env: Rscript: No such file or directory

[vast combine]: Combining results into a single table cat: raw_incl/INCLUSION_LEVELS_ALT3-Xt17-n.tab: No such file or directory cat: raw_incl/INCLUSION_LEVELS_ALT5-Xt17-n.tab: No such file or directory cat: raw_incl/INCLUSION_LEVELS_IR-Xt17.tab: No such file or directory

(The raw_incl folder is also missing the Xt14 and Xt15 versions of these files)

The VTS_LOG_commands file shows that I use this command for alignment: vast-tools align -sp Xt1 -o /mnt/lustre/users/sk1350/tropicalis/vast-tools/vast_out -c 35 -IR_version 2 -stepSize 25 -mismatchNum 2

[mirimia]

Hello Sofia,

I see. Two things:

• could you please open a new Issue? This is no longer related to the topic of the issue. Thanks!
• silly question, but just in case: do you have R installed!? The error "/usr/bin/env: Rscript: No such file or directory" seems very R related to me...