Skip to main content
. Author manuscript; available in PMC: 2017 Mar 1.
Published in final edited form as: Nat Protoc. 2016 Aug 11;11(9):1650–1667. doi: 10.1038/nprot.2016.095

Table 5.

Troubleshooting advice.

Step Problem Possible reason Solution
4 StringTie reports
that it cannot open
the files in the input
mergelist.txt file
The GTF files specified
in mergelist.txt do not
exist, or you did not run
the merge step in the
same directory where
the GTF files were
created.
Either run StringTie in the same
directory as the input GTF files or edit
the mergelist.txt file to include the full
path to each GTF file name
description.
8 You get an error
message that the file
geuvadis_phenodata.csv
cannot be opened.
The directory where you
started your R session
does not contain the
phenotype data.
Before proceeding, you need to tell R
where all the files are. You can do this
with the setwd command. For
example if the directory containing the
chrX_data is your home directory,
then you would use the following R
command:
> setwd(“~/chrX_data”)
You can then run the getwd()
command to make sure that the
directory has been properly set.
9 The Ballgown
function results in an
error that the first
column of pData
does not match the
names of the folders
containing the
ballgown data.
The sample names you
have stored in your
phenotype file do not
match the file names of
the samples you ran
with StringTie.
To make sure that the file names
match the IDs in the phenotype file,
run the following R command:
> all(pheno_data$ids == list.files("ballgown"))
This command should return TRUE.
16 The transcripts and
genes you assembled
(shown in Table 3
and Table 4
respectively) have
different values for
the ‘gene id’ and
‘transcript name’
than what you see in
the protocol here.
The transcript and gene
identifiers assigned by
StringTie (by default
preceded by the string
‘MSTRG’) are
generated in the order
that the bundles of reads
are processed. This
order can differ from
one run of StringTie to
another when multiple
threads are used (with
the –p option).
This is not actually an error. If a
transcript matches a known, annotated
transcript, then the ‘transcript name’
will be consistent from run to run.
Otherwise this field can change when
you re-run the protocol.