search

hansenlab / minfi

Devel repository for minfi
58 stars 70 forks source link

Cell proportions in estimateCellCounts() #66

Open albertod34 opened 8 years ago

albertod34 commented 8 years ago

Hi,

I found some issues understanding the output of minfi's estimateCellCounts(). The thing is that I obtain negative and >1 values. I don't know how to interpret this numbers.

Thanks

Alberto

kasperdanielhansen commented 8 years ago

supply code and output.

albertod34 commented 8 years ago

I have to make a correction, to be more concrete I use projectCellType() function defined inside estimateCellCounts() code. I built my own matrix of reference called sig_matrix. And then, I follow similar steps found in: http://people.oregonstate.edu/~housemae/software/TutorialLondon2014/Houseman_Tutorial_CellMixture_v04.pdf

Thanks for your time Kasper.

sig_matrixNSCLC[1:10,] Cancer Associated Fibroblast Non-Small Cell Lung Cancer cg00813378 0.046362354 0.8804989 cg01959730 0.065177564 0.9539002 cg02283366 0.079840781 0.9266913 cg03596016 0.043515904 0.9343645 cg04021697 0.028698201 0.9551448 cg05039004 0.019765947 0.8709848 cg05047401 0.006049137 0.8508198 cg06756211 0.104746400 0.9375524 cg06776173 0.050983289 0.9663864 cg06962177 0.056461843 0.9545999

tail(sig_matrixNSCLC) Cancer Associated Fibroblast Non-Small Cell Lung Cancer cg02550308 0.03938642 0.9384706 cg06933697 0.06167364 0.9142876 cg08595782 0.05918991 0.9184243 cg06186155 0.04946101 0.9159747 cg22550815 0.07444388 0.9270525 cg06457011 0.10489981 0.9389722

CellType proportions

sig_matrixNSCLC<- as.matrix(sig_matrixNSCLC) cellpropsNSCLC<- minfi:::projectCellType(NSCLCbetas[rownames(sig_matrixNSCLC),], sig_matrixNSCLC)

cellpropsNSCLC Cancer Associated Fibroblast Non-Small Cell Lung Cancer 8942351150_R03C02 1.00959564 -0.003231767 8942351149_R01C01 0.90898458 0.005415805 8942351139_R04C01 1.06441772 0.005632061 8942351148_R06C01 1.05367397 -0.006815078 8942351169_R02C01 0.99591129 -0.007464904 8942351150_R02C01 1.02880817 -0.010452845 6264496054_R06C02 0.91256287 0.013089182 6264509117_R01C02 1.02165966 0.001709435 6285609065_R01C01 0.86466920 -0.001529456 6264509117_R02C02 1.05308210 0.007374339 6264509106_R06C02 1.19877222 0.001678675 6285609059_R03C01 0.88786259 -0.005405449 6055424101_R04C01 -0.25191697 1.064674044 7973201119_R04C01 0.14008469 0.932804513 7970368105_R02C01 -0.15732978 1.023933534 8221924155_R01C01 -0.14709022 1.048732712 8221916070_R03C01 -0.06166177 1.000795772 6055424074_R05C02 0.25895195 0.968434564 7973201120_R06C01 -0.28216584 1.062534257 8221932024_R05C01 -0.26964119 1.036382265 7878191015_R01C01 -0.28573584 1.074431896 8221924068_R02C02 -0.16316676 1.041429168 8221916070_R06C01 -0.08882248 1.059486337 8221924047_R03C02 -0.13437269 1.027898590 7878191064_R03C01 0.24542196 0.958843795 7878191029_R04C01 0.17649157 0.918272913 6055424101_R02C01 0.18638156 0.950201829 7878191066_R01C01 -0.26517653 1.052529972 7970376007_R02C02 0.10530265 0.943261827 7878191066_R02C01 -0.02401698 0.994989271 8221924047_R02C02 -0.18332805 1.041614112 7878191066_R01C02 1.03092434 0.842135800 7970368131_R02C02 0.36031072 0.906871809 7878191064_R01C02 0.01440793 0.971930460 8221916070_R02C01 -0.03749796 1.049436093 7970368089_R04C02 -0.16635433 1.028374466