Skip to content

Latest commit

 

History

History
104 lines (72 loc) · 2.85 KB

naming_rules.md

File metadata and controls

104 lines (72 loc) · 2.85 KB
Raw

Files, directories

  • never ever use space/tabs in your file/directory names
  • never use Unix special characters in file/directory names (!?"'%&^~*$|/{}<>:)

Samples/tissues

  • use few letter prefix for the disease
project prefix
ca. breast cbr
ca. colon cco
ca. prostate cpr
ca. bladder cbl
Alzheimer alz
control ctr
fibroadenoma fad

  • keep the patient/person ID unchanged in all the experimental data.

  • if not created by the in house database system, the patient_id should be: 2 letters researcher (responsible for the patient sample collection) initial followed by 4 numbers.

dk0001
dk9999

pm0543
  • merge prefix (project) with patient_id using one underscore "_"
alz_dk0123
  • provide a code for the tissue
tissue code
blood BL
primary tumor PT
unaffected margin UM
unaffected margin 99 UM99
  • merge with project code & patient_id using underscore
cbr_ak0125_PT
  • if not provided with a sample code id from the in house database:

provide the sample number: 0 for the first sample of that type (i.e. the tube #1 of blood used for DNA extraction) 1, 2 etc for the subsequent material preparations from the given patient-sampletype combo

cbr_ak0125_PT_0

cbr_ak0125_PT_1
  • in case of repeating the experiment, i.e. re-sequencing the same sample, lets add another number. To be discussed
Hello,

currently I started to use the following system:

cbrako_fix034_PT_0_9
cbr  = > ca breast (major disease prefix)
ako =>  specific project / sample source if applicable. Can be xyz for all samples stored here
fix034 => internal to that project patient id
PT => tissue type
_0_9 => biological-replicate_technical-replicate

I also used PT_13_0 for the tumor in the same patient after remission.

Above system is a quick fix for tissue samples stored in tubes labeled with non-consistent patient id's followed by generating ~10k sequencing files with groups of 4-8 with ~identical patient/tissue identifiers (fix034_PT.result,  fix034PT.result, fix034-PT.result,  fix034PT.result, etc).

Since we will get the unique ids for all the tissue samples, we can do:
a) cbrako_fix034_PT_xd2f_9 (where xd2f is some unique string to identify tumor samples from fix034 patient
b) enumerate tumor/blood samples from a given patient 
xd2f 0 
jk5p 1 

keep the  cbrako_fix034_PT_0_1
c) go for an overkill
cbrako_fix034_PT_0_1.xd2f.

where xd2f will be id for DNA/RNA prep.

I am concatenating labels for a various processing steps for each of the data files:
cbrako_fix039_PT_0_2.rna.mrgd_4.clump.r12.star_hg38p13.md.bam

to minimize the chance of screwing things and be able to figure out things in a week or in a year from now. But it is a non-trivial problem.

This is just to keep you posted before the tomorrows meeting/save some time.