Each node of cluster should call map.r by using
R CMD BATCH --slave '--args rdata_file="<preprocessed inputs>.rdata" chtc_process_number="$1"' map.r
- rdata_file (see make_rdata.r for details):
- traits: List of traits. trait_index argument specifies which trait to map with
- mapping_jobs: List of mapping_job objects, each containing specifics of mapping to be done
- Fields (per mapping_job object):
- cross: R/QTL cross object containing genotype and phenotype information
- sex: 'female', 'male', or 'hetero'
- log_status: "logged" or "not-logged" specifies if log or raw trait values should be used
- covariate_type: 'additive', or 'interactive' (see below for difference)
- covariate_trait: name of the trait to use in covariate mapping
- Fields (per mapping_job object):
- The filenames indicate if males, females, or both males and females ('hetero') were used in the mapping.
- The filenames also indicate if the log or raw values were used ('logged' vs 'not_logged')
- When both males and females were mapped together, the filename indicates if sex was used as an additive covariate ('the average phenotype is allowed to be different in the two sexes, but the effect of the putative QTL is assumed to be the same in the two sexes.'-Karl Broman) or as both an additive covariate and an interactive covariate (allows QTL to be different in each sex)
- Finally, there are 4 types of files generated by each mapping:
- .txt file has LOD scores for all markers
- .log file has marker with highest LOD score for each chromosome (if summary='TRUE' in map.sh)
- .pdf file has the LOD plots with all markers and all chromosomes (if pdf='TRUE' in map.sh) (.pdf used instead of .png b/c chtc nodes can't make .pngs out of the box)
- .rdata file has the R/QTL scanone object, which can be loaded back (if rdata='TRUE' in map.sh) into R to make additional plots