Thursday 11/21/13

9:30a – 11:15p

Meetings

  • working on slides for chat with Mike.
  • meeting with Mike Levine
  • Levine seminar at HMS

Genomics

  • installed cufflinks on local scratch folder
  • successfully (? or not) ran on Rep1 data from ENCODE.
  • All FPKM and coverage values are 0.
  • Not clear what isoforms are what
  • Output header does not fully match description from cufflinks website.

Troubleshooting round 1

  • convert to SAM, then run cufflinks with SAM file target.

Odyssey 2.0 New format

  • New Module List for CentOS6: https://rc.fas.harvard.edu/module_list/#Centos6
  • System has new job manager
  • will need to rewrite all the matlab script for interacting with the jobs list, bjobs and bsub are all extinct.
  • Caltech hESC data runs without error after converting to SAM. SAM file was highly abreviated due to out of disk space error. Only the first few genes seem to have run correctly.
  • CSHL A549 data doesn’t seem to be running correctly — no output files are appearing.
  • Downloading cytosolic A549 data in place of whole cell data. Maybe the header on these files won’t be screwed up.
  • ah ha! files finally appearing now. Maybe this might work…

New Odyssey commands

sbatch batchfilename.bat launches job specified in batchfile

 #!/bin/bash 

 #SBATCH -n 4  #Number of cores 

 #SBATCH -t 1000  #Runtime in minutes 

 #SBATCH -p general  #Partition to submit to 

 #SBATCH --mem=1000 #Memory per node in MB (see also --mem-per-cpu) 


 cufflinks -o ./Genomics/A549/ -G ./Genomics/gencode.v18.annotation.gtf ./Genomics/wgEncodeCshlLongRnaSeqA549CytosolPapAlnRep3.bam > output.txt 2> errors.txt

Note the line breaks must be in unix format. Can specify unix style line endings in Notepad++

Other useful commands

sacct -o MaxRSS -j jobID# returns the maximum memory used by this job in Kb. Useful in choosing the correct memory limits for launching the next job. Note request requiring large amounts of memory have a much harder time fitting in on the queue.

sacct -j jobID# returns the run status (PENDING/RUNNING/COMPLETED/FAILED) for the indicated job.

sacct without an argument returns all jobs (launched in this session?)

STORM

  • attempting to setup STORM2
  • back-optics still substantially out of alignment, outside maximal range of tuning mirrors. Failed to get back in alignment.
This entry was posted in Summaries and tagged , , . Bookmark the permalink.