无NA值运行Slingshot仍报'Error in if (!useNames)'错误求助

问题背景 #

在snRNA-seq数据上运行Slingshot工具时，执行以下代码持续报错，已排查确认数据无NA值、聚类标签匹配，更换聚类标签后仍出现相同错误：

sce <- slingshot(sce, clusterLabels = sce$seurat_clusters, reducedDim = "UMAP")
Error in if (!useNames) { : missing value where TRUE/FALSE needed

调用栈信息：

> sce <- slingshot(sce, clusterLabels = sce$slingshot_clusters, reducedDim = "UMAP")
Error in if (!useNames) { : missing value where TRUE/FALSE needed

Enter a frame number, or 0 to exit   

 1: slingshot(sce, clusterLabels = sce$slingshot_clusters, reducedDim = "UMAP")
 2: slingshot(sce, clusterLabels = sce$slingshot_clusters, reducedDim = "UMAP")
 3: .local(data, clusterLabels, ...)
 4: slingshot(data = rd, clusterLabels = cl, reducedDim = NULL, start.clus = start.clus, end.clus = 
 5: slingshot(data = rd, clusterLabels = cl, reducedDim = NULL, start.clus = start.clus, end.clus = 
 6: .local(data, clusterLabels, ...)
 7: getLineages(data, clusterLabels, reducedDim = reducedDim, start.clus = start.clus, end.clus = en
 8: getLineages(data, clusterLabels, reducedDim = reducedDim, start.clus = start.clus, end.clus = en
 9: getLineages(data = data, clusterLabels = clusWeight, ...)
10: getLineages(data = data, clusterLabels = clusWeight, ...)
11: .local(data, clusterLabels, ...)
12: createClusterMST(x = X[use, , drop = FALSE], clusters = clusterLabels[use, , drop = FALSE], outg
13: createClusterMST(x = X[use, , drop = FALSE], clusters = clusterLabels[use, , drop = FALSE], outg
14: .local(x, ...)
15: FUN(x, clusters)
16: .rowstats_w(DelayedArray::DelayedArray(x), group, FUN = DelayedMatrixStats::colWeightedMeans)
17: FUN(x, w = group[, i], ...)
18: FUN(x, w = group[, i], ...)
19: .local(x, w, rows, cols, na.rm, ..., useNames = useNames)
20: .smart_seed_dispatcher(x, generic = MatrixGenerics::colWeightedMeans, blockfun = .DelayedMatrix_
21: candidate(S, ...)
22: matrixStats::colWeightedMeans(x, w = w, rows = rows, cols = cols, na.rm = na.rm, ..., useNames =

解决方法 #

• 更新相关依赖包：该错误常因matrixStats、slingshot或DelayedArray版本不兼容导致，优先更新这些包：

  update.packages(c("slingshot", "matrixStats", "DelayedArray"), ask = FALSE)
  

• 转换降维矩阵为普通矩阵：如果UMAP降维结果是DelayedArray类型，可能引发参数传递问题，转换为常规矩阵后重试：

  reducedDim(sce, "UMAP") <- as.matrix(reducedDim(sce, "UMAP"))
  sce <- slingshot(sce, clusterLabels = sce$seurat_clusters, reducedDim = "UMAP")
  

• 规范聚类标签类型：确保聚类标签是因子或字符型，避免数值型可能引发的隐式转换问题：

  sce$seurat_clusters <- as.factor(sce$seurat_clusters)
  # 再次确认无NA值
  table(is.na(sce$seurat_clusters))
  

• 手动指定发育起始/结束聚类：自动推断 lineage 时可能出现异常，手动指定起始和结束聚类缩小计算范围：

  # 替换为实际的聚类ID
  sce <- slingshot(sce, clusterLabels = sce$seurat_clusters, reducedDim = "UMAP", 
                   start.clus = "0", end.clus = c("4", "6"))
  

• 简化数据测试：取少量细胞子集运行，排查是否因数据量过大或个别异常细胞导致：

  set.seed(123)
  sce_sub <- sce[, sample(ncol(sce), 1000)]
  sce_sub <- slingshot(sce_sub, clusterLabels = sce_sub$seurat_clusters, reducedDim = "UMAP")
  

内容的提问来源于stack exchange，提问作者Toon Leroy