I tried to do Enrichment Analysis of Arabidopsis thaliana by using enrichGO with R. But the result is empty. What caused this result?

> ego <- enrichGO(syn_block1_gene, OrgDb = "org.At.tair.db", keyType = "TAIR",ont = "BP")
> ego
#
# over-representation test
#
#...@organism    Arabidopsis thaliana 
#...@ontology    BP 
#...@keytype     TAIR 
#...@gene    chr [1:754] "AT1G55230" "AT1G55240" "AT1G55380" "AT1G55390" "AT1G55420" ...
#...pvalues adjusted by 'BH' with cutoff <0.05 
#...0 enriched terms found
#...Citation
  Guangchuang Yu, Li-Gen Wang, Yanyan Han and Qing-Yu He.
  clusterProfiler: an R package for comparing biological themes among
  gene clusters. OMICS: A Journal of Integrative Biology
  2012, 16(5):284-287 

> dotplot(ego)
> head(ego)
[1] ID          Description GeneRatio   BgRatio     pvalue      p.adjust   
[7] qvalue      geneID      Count      
<0 rows> (or 0-length row.names)
> as.data.frame(ego)[1,]
     ID Description GeneRatio BgRatio pvalue p.adjust qvalue geneID Count
NA <NA>        <NA>      <NA>    <NA>     NA       NA     NA   <NA>    NA

And I'm sure the data of syn_block1_gene is correct.

The absence of errors indicate that the analysis as such went fine. The fact that no enriched GO terms were found thus indicate that there are none! At least, when using the default cutoff values (since you didn't specify them yourselves; see 4th line of your object ego):

#...pvalues adjusted by 'BH' with cutoff <0.05

So to double-check your analysis went fine, I would rerun enrichGO(), but then without applying any significance cutoff. Thus by setting/including the arguments pAdjustMethod = "none", pvalueCutoff = 1, and qvalueCutoff = 1.

I recently posted a similar answer: "No genes can be mapped...." using enrichGO in clusterProfiler