(1) "Error waiting for container: invalid character 'u' looking for beginning of value" (2) "Could not execute because the application was not found or a compatible .NET SDK is not installed." #212
Comments
|
Hmm, I've never seen that one before. Waiting for 'u' is a particularly odd error message, too! It seems like this is a Docker issue that's still open: docker/for-mac#5139, dealing with the VM freezing. Have you gotten this error multiple times? |
|
Yes, it comes and goes... when i restart, it seems to get stuck in following steps... never made it to the end of this data, I can probably try to ditch the docker and run natively? Is there some "dryrun" output of spritz which I can go one step at a time? |
|
Sorry about Docker being flakey. That's frustrating. For the time being, I'd recommend using the commandline version, https://github.com/smith-chem-wisc/Spritz/wiki/Spritz-commandline-usage. You should be able to specify the directory you're using for the analysis in the "analysisDirectory" config.yaml specification. |
|
Thanks @acesnik , just to confirm before i proceed, I see there a config.yaml which I am guessing is created by the GUI? I am wondering if I need to change the |
|
The The snakemake command you listed looks good. |
|
Thanks @acesnik for looking into this 👍🏼 but I am not sure where to run the snakemake command from? I tried to find the makefile or the workflow without success? Any ideas where it might be or which directory to initiate the command from? |
|
Try searching for the |
|
In other words, assuming your git clone is named |
|
Looks like but htop shows stuff are running? |
|
Oh, I see, I think it started earlier than you were before because the resources that were downloaded into the docker container (genome, gene model, etc) were downloaded again, so the timestamps of those resources are later than the previous place you were at... Regarding the files that are write-protected, I wonder if there are any hanging docker containers. You can see if there are any still running with I think the best thing to do now would be to just to let it run, or to start it again with the |
|
I reboot the machine, so docker is out, alteast that's what I get when BTW the last command crashed with message |
|
Okay! Hoping for the best! Thanks for the patience with this one! |
|
Did the rest of the run go okay for you? |
|
It crashed with that "Write-protected" log.2.txt error , tried chown -R which didn't work, probably need to restart the machine but waiting for some other work to finish first... is there anyways to go over this without restarting? |
|
Is it possible to remove these files manually? |
|
Now its blaming "Write-protected output files for rule hisat2_align_bam_fq: |
|
I've never seen this before. Are the other snakemake processes still hanging around for some reason? |
|
|
|
There are couple of other process but i think these are unrelated? |
|
Hmm, yeah that looks like it's from a pipeline named |
|
so you think conda envs are cross-contaminating? |
|
No, I don't think they're colliding. I really don't know why those output files are write protected. That's pretty mysterious! I do think restarting after the other runs finish is a good idea. |
|
Yes that is the plan, will get back once through 👍🏼 |
|
I eventually ended up deleting the whole |
|
Are you using version 0.2.4? That looks familiar from errors we were getting in v0.2.3. |
|
|
|
Just running |
|
That is the latest commit. Thanks for checking! I'll look into this more later today. |
|
Sorry for the delay on this. I took a look but didn't get very far. I'm traveling now, so I'll be able to take a closer look in a couple weeks. |
|
Looks like it went past that issue but it does say |
|
Interesting. It looks like some of the transcript assemblies are failing to build. Could you send me the log files at |
|
I don't think the issue is from bumping the genome version. I've seen these issues before when the aligned read counts are low for some reason. |
|
You could also look at some of the log files from aligning these files to see if that's the case, e.g. if there are low read counts. In a typical experiment, I would expect >80% or >90% of reads to be aligned. I've seen these types of errors when there are <20% aligned, for example, which points to an alignment issue. |
|
If spritz cannot detect isoforms with stringtie, I'd recommend just finishing the pipeline with the |
|
OK, i have disabled isoform call and reran the workflow but it crashed the WSL itself... |
|
Thanks for giving that a try! Could you tell me more about the WSL crash? If I remember correctly, you have 64 GB of RAM, so that's probably not the issue... Thanks for the information about the *.sorted.gtf.log files being empty. Please check on the files at |
|
Yes i have 64GB RAM free but it looks like snakemake directory lock was the issue as a rerun after invoking with snakemake with --unlock Thanks be to you @acesnik for making&sharing, may the force be with you 👍🏽 |
|
Okay, great! No problem! I'm going to close this issue. Feel free to reopen or open a new one if you run into anything new! |
|
It went fine for a while but then crashed |
|
It looks like the drive where spritz is located may have run out of storage space. I think this happens when it's writing temporary files and the temporary file location (../resources/tmp) runs out of space. Could you share /home/animeshs/rnAGS/variants/Homo_sapiens.GRCh38.100.spritz.log, as well? |
|
I think drive space should not be the issue as /mnt/z has the data... looking into the log file you asked for |
|
The latter issue in Homo_sapiens.GRCh38.100.spritz.log was fixed here: #217 Could you try doing |
|
Is your With it there where there's plenty of space, I really don't know why a file is closing in the middle of the SplitNCigarReads tool execution. Could you please check whether temporary files being saved to I just double checked the option for specifying the temporary directory ( |
|
OK, i have pull/move and re-running the pipeline but after this it keeps crashing at 1st step? Should i just redo or something else i can looking into for saving the work so far? |
|
I'm not sure what's going wrong based on that output. Is it producing any log files? I also wonder whether the analysis directory is specified as an absolute path in the config file. |
|
Looks like it is getting stuck and conda activation stage, below is the one i re-ran and just cancelled, it was running since yesterday... Below is the config, it is a link to absolute pathj, should it be relative? |
|
You could try deleting the |
|
Looks like it config.zip finally worked 2021-08-13T153310.143502.snakemake.log @acesnik 👍🏽 I am guessing that following are the results for the variant calling? If so, are the following numbers of protein sequences look reasonable to you? The with decoys numbers are confusing me though, like 152804 is off by 26194 if the reverse is included like 2*63305 or does it contain the variants too? The number difference is 29709 for the last two... How are the variants being encoded in the fasta files? Is there a straightforward within Spritz way to summarize those? |
|
These two files are actually the results of different programs, which should probably be clearer in the filenames. The first ones ( If you check |
|
If you want to use the FASTA file with variants but without decoys, I would recommend selecting |
|
Great @acesnik , 152804/2=>76402 matches perfectly though 88340 and 88345 don't, does it mean there are only 5 variants? do you think the results are fine in general? Should i go forward with /home/animeshs/rnAGS/final/combined.spritz.snpeff.protein.withdecoys.fasta* none the less? |
|
Thanks, looks like you replied in between @acesnik . So i am guessing the variants are being encoded within the ^mz_* ? How to extract them and confirm, any way within Spritz? |
|
Oh, that's interesting about the extra decoys. There is a process of reversing the variants in decoy generation, and that might change the length and thus the filtering based on having proteins >7 AAs. I'll look into that. The decoy discrepancy is pretty small, so it might not have the biggest effect, but like I mentioned, you could select just the targets with All target proteins are encoded with "^>mz", and all decoys are encoded with "^>rev_mz". Variants will have the tag "variant:" in the header, so to check for the count of proteins with variants, I'd recommend doing |
|
In typical human experiments, I usually find about a quarter of the entries have variants. |
|
Awesome @acesnik 👍🏽 Looks like there are about 15000 variants which is less than expected then? Probably the read depth was not enough, where to check for the mapping metrics and play with thresholds used for variant-calling? Is there some easy to configure within the config.yaml? |
|
That's right around where I would expect. Probably 15-30% of entries or something. I wouldn't try changing any of the thresholding, personally. It's actually probably more than 15000 variants, since multiple are encoded per protein. You can check on some other figures in the output of SpritzModifications at The first one also tells you the number of variant containing proteins (hopefully the same number as you saw above), the number of unique variants on those proteins, and the number of unique variants by type: missense, frameshift, deletion, etc. |
|
You can also take a step back from there and look at the numbers of variants detected before applying them to proteins by opening |
|
The numbers seem to be matching withmods.log I will check the HTML too, thanks a lot @acesnik 👍🏽 |
Docker is running though... below is the full log
The text was updated successfully, but these errors were encountered: