使用MetaIntegrator的runMetaAnalysis函数筛选探针后报错求助

问题场景 #

使用MetaIntegrator包的runMetaAnalysis函数对特定探针子集筛选后运行，出现探针不匹配及输入对象错误。

原代码 #

geo.objects <- c("GSE51387",  "GSE64660")
geo.object <- getGEOData(geo.objects, formattedNames = geo.objects, qNorm = FALSE)
probeIDs_of_interest <- c("A_51_P385099", "A_51_P283759", "A_52_P32733", 
                          "A_52_P569539", "A_52_P582969", "A_51_P482552", 
                          "A_52_P229471", "A_52_P249424", "A_52_P638895", 
                          "A_51_P458168", "A_52_P436628")


geo.object$originalData$GSE51387$keys <- geo.object$originalData$GSE51387$keys[names(geo.object$originalData$GSE51387$keys) %in% probeIDs_of_interest]

geo.object$originalData$GSE51387$expr <- geo.object$originalData$GSE51387$expr[probeIDs_of_interest, ]

geo.object$originalData$GSE64660$keys <- geo.object$originalData$GSE64660$keys[names(geo.object$originalData$GSE64660$keys) %in% probeIDs_of_interest]

geo.object$originalData$GSE64660$expr <- geo.object$originalData$GSE64660$expr[probeIDs_of_interest, ]
metaobject <- runMetaAnalysis(geo.object)

错误信息 #

metaobject <- runMetaAnalysis(geo.object)

Warning: The probes listed in expr and the probes listed in keys do not match.
Warning: The probes listed in expr and the probes listed in keys do not match.
Warning: $originalData contains some datasets that either are null, the wrong type, or have mismatched samples. These are the datasets: GSE51387 GSE64660
Error in runMetaAnalysis(geo.object) : Error in the input object!

报错原因 #

1. 部分目标探针在某个数据集中不存在，强行筛选后expr会出现空行/NA，而keys只保留存在的探针，导致两者探针集合不匹配。
2. 直接用probeIDs_of_interest筛选expr时，若探针不存在会生成NA行，破坏数据结构，被runMetaAnalysis判定为无效数据集。

解决方案 #

修正后的代码 #

geo.objects <- c("GSE51387",  "GSE64660")
geo.object <- getGEOData(geo.objects, formattedNames = geo.objects, qNorm = FALSE)
probeIDs_of_interest <- c("A_51_P385099", "A_51_P283759", "A_52_P32733", 
                          "A_52_P569539", "A_52_P582969", "A_51_P482552", 
                          "A_52_P229471", "A_52_P249424", "A_52_P638895", 
                          "A_51_P458168", "A_52_P436628")

# 处理GSE51387：仅保留数据集中存在的目标探针
valid_probes_gse51387 <- intersect(probeIDs_of_interest, names(geo.object$originalData$GSE51387$keys))
geo.object$originalData$GSE51387$keys <- geo.object$originalData$GSE51387$keys[valid_probes_gse51387]
geo.object$originalData$GSE51387$expr <- geo.object$originalData$GSE51387$expr[valid_probes_gse51387, ]

# 处理GSE64660：仅保留数据集中存在的目标探针
valid_probes_gse64660 <- intersect(probeIDs_of_interest, names(geo.object$originalData$GSE64660$keys))
geo.object$originalData$GSE64660$keys <- geo.object$originalData$GSE64660$keys[valid_probes_gse64660]
geo.object$originalData$GSE64660$expr <- geo.object$originalData$GSE64660$expr[valid_probes_gse64660, ]

# 提前检查：避免数据集无有效探针导致后续报错
if (length(valid_probes_gse51387) == 0 || length(valid_probes_gse64660) == 0) {
  stop("某个数据集没有匹配的探针，请核对探针ID是否正确")
}

metaobject <- runMetaAnalysis(geo.object)

关键修正点 #

• 用intersect获取每个数据集与目标探针的交集，确保只筛选存在的探针，避免expr和keys出现不匹配。
• 统一用交集探针筛选keys和expr，保证两者的探针顺序、数量完全一致。
• 添加前置检查，提前发现无有效探针的数据集，避免后续分析报错。

内容的提问来源于stack exchange，提问作者Nesma M