Christopher Hann-Soden
Christopher Hann-Soden
I've run into this (or a similar) issue attempting to install Salmon on the UC Berkeley HPC cluster. Iconv was present within one of our Python installs, but that didn't...
We dug into this on the Sleuth issue, and eventually passed it on as an issue in rhdf5. My workaround was to create a Sleuth Docker container. See https://github.com/pachterlab/sleuth/issues/175#issuecomment-397451628
Thank you for your prompt reply! It does indeed fail at the first check - obs_raw is entirely empty.
The read_kallisto function doesn't look like it's succeeding. e.g.: read_kallisto("wt/W795_22C_L_corrected_kallisto", read_bootstrap = FALSE) kallisto object transcripts: 0 bootstraps: 0
The files don't look empty to me. The tsvs look sane enough, and the h5s are several MB. The error logs for the kallisto runs all looked normal enough to...
The fields of test_kal are entirely NULL. Weird. I'll run a checksum on the original files, maybe try running Kallisto over again to see if I get something different.
FYI, the checksums check out, so if they got corrupted it was on my serve filesystem somehow.
I repeated the run with v.0.44.0 to the same effect. Should I open up a bug for Kallisto? I can provide you with additional sample data for the 0.44.0 run.
@pimentel Yes, I am still getting the exact same errors. I put the new files in same path as before, and when I do ``` test_kal
Hey, thanks for helping me trouble-shoot this. Good to narrow down the source of the problem - I may try a different machine to see if I can replicate. The...