search

Teichlab / tracer

TraCeR - reconstruction of T cell receptor sequences from single-cell RNAseq data
Other
124 stars 48 forks source link

Tracer Summarise failed when plotting reconstructed lengths #40

Closed mn120110d closed 7 years ago

mn120110d commented 7 years ago

Hi Mike,

We wanted to test TraCeR with files from the link: https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE45428. There are provided three samples: Jurkat Cells 216, 236 and 245. The samples are provided in SRA format, so we used SRA Toolkit Fastq-Dump available here: https://github.com/ncbi/sra-tools for generating fastq files for these paired end reads. First, we ran TraCeR Assemble for fastq files generated for Jurkat Cells 216. In attachment you can find the results of the assemble phase. Than we tried to run TraCeR Summarise, but we got an exception on line sns.distplot() in file tasks.py. Here is also the output of the program that produced this error: Tracer Summarise dump.txt

When we commented the line sns.distplot() in tasks.py, the program executed fine. Is it possible for you to handle this exception in the code, so the program can continue with further execution without plotting reconstructed lengths?

Thank you! Tamara and Nevena

Results of the assemble phase: (Species set to Hsap for both assemble and summarise) cell_line_216.zip

boombard commented 7 years ago

Hi

I can't reproduce your error with your assembly output. I think it may be a problem that was resolved in an update of the seaborn library, when there is only one point for a kde plot.

Could you tell me which version of seaborn you currently have installed with pip show seaborn? And how have you installed tracer?

Here are the figures that created the error you are seeing:

reconstructed_lengths_TCR_A.pdf reconstructed_lengths_TCR_B.pdf

mn120110d commented 7 years ago

Hi,

Thank you for your reply and for the figures. We have installed seaborn version 0.7.1. Other libraries related to seaborn have versions: numpy 1.12.1 and scipy 0.19.0. We installed tracer v0.5.1 like this: git clone https://github.com/Teichlab/tracer tracer cd tracer/ git reset --hard e109bfbe25c64c9aedf9a540a7592d11162e2493

We are running it on Docker image, on Ubuntu 14:04. We have installed all other necessary dependencies and when we run tracer with your test data for mouse, we get correct results.

mn120110d commented 7 years ago

Hi,

We found the solution to this problem. It turned out that we didn't have installed statsmodels. Now that we have installed it, we get the expected output. Thank you for your help!