############################################################





Version 1 of MTG is not actively maintained anymore.
For the latest version of MTG please visit Sang Hong Lee's website at
https://sites.google.com/site/honglee0707/mtg2





############################################################ 


  MTG (Multi-trait GREML and GBLUP program) 
  version1.3
  SH Lee (Jul/2014)

  Citation:
  Maier, R., et al. Joint analysis of psychiatric disorders increases accuracy of risk prediction for schizophrenia, bipolar disorder and major depression disorder. The American Journal of Human Genetics 96, 283-294 (2015)


  mtg1.3     : compiled with intel ifort (possible to parallel e.g. -thread k)
  rtmx_frq-vx: estimating mean(x) and var(x) (x is 0,1,2 SNP coefficient)
  snp_blup1.5: converting individual BLUP to SNP BLUP or the otherway around

  NOTE: when scaling x, we use var(x) rather than 2p(1-p) (p is RAF)


  <2 traits model>
  -To get variance components (-out) and breeding values (-bv),
  ./mtg1.3 -p toy.fam -d toy.phen -g toy.rtmx -out toy.2t.ascm -mod 2 -bv toy.2t.bv -cove 1

  - To get SNP BLUP for the first trait,
  awk '$1==1 {print $2}' toy.2t.bv.py > tmp1
  ./snp_blup1.5 toy tmp1 a toy.2t.snpv_for_t1

  - To get individual BLUP for the first trait in the data 'toy' (looks for toy.bed, toy.bim and toy.fam),
  ./snp_blup1.5 toy toy.2t.snpv_for_t1 b toy.2t.gbv_for_t1

  Here, the toy dataset has been used for both conversion, so the BLUP for the first trait in toy.2t.bv and toy.2t.gbv_for_t1 are identical. It is usuallt recommended to use an independent validation dataset by replacing the second argument, e.g. 
./snp_blup1.5 your_plink_file toy.2t.snpv_for_t1 b your_plink_file.gbv_for_t1




  <3 traits model>

  - To get variance components (-out) and breeding values (-bv),
  ./mtg1.3 -p toy.fam -d toy.phen -g toy.rtmx -out toy.3t.ascm -mod 3 -bv toy.3t.bv -cove 1

  - To get SNP BLUP for the first trait,
  awk '$1==1 {print $2}' toy.3t.bv.py > tmp1
  ./snp_blup1.5 toy tmp1 a toy.3t.snpv_for_t1

  - To get SNP BLUP for the third trait,
  awk '$1==3 {print $2}' toy.3t.bv.py > tmp3
  ./snp_blup1.5 toy tmp3 a toy.3t.snpv_for_t3

  - To get individual BLUP for the third trait in the data 'toy' (which could be replaced with any validation set for which prediction should be made)
  ./snp_blup1.5 toy toy.3t.snpv_for_t3 b toy.3t.gbv_for_t3

  This gives the same values between trait 3 in toy.3t.bv and toy.3t.gbv_for_t3, because the toy dataset has been used both times (i.e. same genotypes, same allele frequencies, same SNP effects and so on). It is usually recommended to use an independent validation dataset by replacing the second argument, e.g. 
./snp_blup1.5 your_plink_file toy.3t.snpv_for_t3 b your_plink_file.gbv_for_t3


  <with binary GRM>
  ./mtg1.3 -p toy.fam -d toy.phen -g toy.rtmx -out toy.2t.ascm -mod 2 -bv toy.2t.bv -cove 1 (with text GRM file)
  ./mtg1.3 -p toy.fam -d toy.phen -bg toy.grm.bin -out toy.2t.ascm -mod 2 -bv toy.2t.bv -cove 1 (with binary GRM file (GCTA format))



  Please see more detail in manual20150708.docx