Error when trying to check Cell-Cycle impact in PCA

[Ima97]

Nice to meet you.
I am new to bioinformatics.
Thank you for taking a look at this article.

I tried to do a Single cell RNA-Seq analysis using Seurat.

However, when I tried to check the effect of Cell-Cycle in PCA, I got the following error.

Error in irlba(A = t(x = object), nv = npcs, ...) : 
  max(nu, nv) must be positive
In addition: Warning message:
In PrepDR(object = object, features = features, verbose = verbose) :
  The following 97 features requested have not been scaled (running reduction without them): Mcm5, Pcna, Tyms, Fen1, Mcm7, Mcm4, Rrm1, Ung, Gins2, Mcm6, Cdca7, Dtl, Prim1, Uhrf1, Cenpu, Hells, Rfc2, Polr1b, Nasp, Rad51ap1, Gmnn, Wdr76, Slbp, Ccne2, Ubr7, Pold3, Msh2, Atad2, Rad51, Rrm2, Cdc45, Cdc6, Exo1, Tipin, Dscc1, Blm, Casp8ap2, Usp1, Clspn, Pola1, Chaf1b, Mrpl36, E2f8, Hmgb2, Cdk1, Nusap1, Ube2c, Birc5, Tpx2, Top2a, Ndc80, Cks2, Nuf2, Cks1b, Mki67, Tmpo, Cenpf, Tacc3, Pimreg, Smc4, Ccnb2, Ckap2l, Ckap2, Aurkb, Bub1, Kif11, Anp32e, Tubb4b, Gtse1, Kif20b, Hjurp, Cdca3, Jpt1, Cdc20, Ttk, Cdc25c, Kif2c, Rangap1, Ncapd2, Dlgap5, Cdca2, Cdca8, Ect2, Kif23, Hmmr, Aurka, Psrc1, Anln, Lbr, Ckap5, Cenpe, Ctcf, Nek2, G2e3, Gas2l3, Cbx5, Cenpa

I will also include the code for R at that time.

#Loading cell cycle marker list
s.genes <- sapply(cc.genes.updated.2019$s.genes, function(x){paste0(substr(x,1,1), tolower(substr(x,2,nchar(x))))})
g2m.genes <- sapply(cc.genes.updated.2019$g2m.genes, function(x){paste0(substr(x,1,1), tolower(substr(x,2,nchar(x))))})

#Checking the impact of Cell-Cycle on PCA
# Scaling the entire data
data.integrated <- ScaleData(object = data.integrated, features = rownames(data.integrated), verbose = FALSE)
# PC
data.integrated <- RunPCA(data.integrated, features = c(s.genes, g2m.genes))

We apologize for the inconvenience, but if anyone knows the cause of this error or how to resolve it, please let us know.

[gilstel]

I get the same error
Is there a solution?

[Gilgamesh-521]

I didn't get the error but a very similar warning (it can be regarded as an arror cuz nothing left for PCA after removing the warning genes). Has anyone found the reason?

  The following features were not available: MCM5, FEN1, RRM1, UNG, GINS2, MCM6, CDCA7, DTL, PRIM1, UHRF1, HELLS, RFC2, RPA2, NASP, RAD51AP1, GMNN, WDR76, SLBP, CCNE2, UBR7, POLD3, MSH2, ATAD2, RAD51, EXO1, TIPIN, DSCC1, BLM, CASP8AP2, USP1, POLA1, CHAF1B, NDC80, CKS1B, TMPO, TACC3, AURKB, ANP32E, KIF20B, RANGAP1, NCAPD2, PSRC1, LBR, CKAP5, CTCF, G2E3, GAS2L3, CBX5, CENPA.
Warning in irlba(A = t(x = object), nv = npcs, ...) :
  You're computing too large a percentage of total singular values, use a standard svd instead.
Warning in irlba(A = t(x = object), nv = npcs, ...) :
  did not converge--results might be invalid!; try increasing work or maxit```

[rharao]

Replying on an old post in case this helps anyone who has the problem now:

check the effect of Cell-Cycle in PCA

data.integrated <- RunPCA(data.integrated, features = c(s.genes, g2m.genes))

In this function call, by setting features = c(s.genes, g2m.genes), you're calculating the PCA using only the cell cycle features as data for each cell; no other biological information is incorporated. Thus, the PCA will necessarily be exclusively influenced by these cell cycle features.

I assume this is not what you want. Here's what you could do instead, starting after the scaling step:

• Calculate cell cycle scores for each cell: data.integrated <- CellCycleScoring(data.integrated, s.genes, g2m.genes)
• Calculate the PCA on the scaled features: `data.integrated <- RunPCA(data.integrated)
• Now, make some visualization that will allow you to identify if any PCs are strongly related to cell cycle state. There are a few ways you can do this.
  • VizDimLoadings or DimHeatmap can show if any of the top few PCs are strongly loaded with cell cycle features.
  • PCAPlot, or VlnPlot could also be applied for this; use the PCs as the coordinate variables and cell cycle phase assignment as grouping variables.
  • Get the relevant data with Embeddings(data.integrated) and data.integrated[[]][c("S.Score", "G2M.Score")] and plot them however you want.