Singular Matrix in MQM

24 views
Skip to first unread message

Nathaniel Burner

unread,
Mar 8, 2024, 12:35:22 PMMar 8
to R/qtl discussion
Hi Karl,

I am attempting to perform MQM and am getting a singular matrix error when calling the mqmscan command. My code is as follows:

rils = jittermap(rils) # jitter markers

rils = convert2riself(rils) # MQM can only handle advanced RILs. Hets are removed

rils_imp = fill.geno(rils, map.function = "kosambi") # impute missing genotypes
geno.image(rils_imp, col = c("white","tomato","dodgerblue")) # red for AA, blue for BB


summary(rils_imp)
RI strains via selfing No. individuals: 188 No. phenotypes: 14 Percent phenotyped: 100 100 100 100 100 100 100 100 100 100 100 100 100 100 No. chromosomes: 20 Autosomes: 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 Total markers: 1548 No. markers: 59 105 75 80 61 93 70 68 56 72 49 97 117 83 76 65 67 97 102 56 Percent genotyped: 100 Genotypes (%): AA:52.1 BB:47.9

phenos = phenames(rils_imp)
pheno = "CW_combined_WYO"

mqm = mqmscan(rils_imp, pheno.col = 3, step.size = 2.0, window.size = 25,
              cofactor.significance = 0.05)
Singular matrixError in mqmscan(rils_imp, pheno.col = 3, step.size = 2, window.size = 25, : Fatal error

plot.map(pull.map(rils_imp),est.map(rils_imp))


I am not sure what the issue here is and I have not found any solutions on the discussion board.

Thanks,
Nathaniel
Rplot01.tiff
Rplot.tiff

Nathaniel Burner

unread,
Mar 8, 2024, 12:48:50 PMMar 8
to R/qtl discussion
Removing duplicate markers resolved this issue. This is an F5 population and I removed duplicate markers when making the initial map. Since MQM does not take F5s, I converted to a RISELF, which converted some of the marker calls and therefore some markers became duplicates of others.
Reply all
Reply to author
Forward
0 new messages