Picard CollectRnaSeqMetrics error

Question

[juncheng]

I want to get mapping statistic from a mapped bam file and want to use CollectRnaSeqMetrics.

I have a gtf file which is downloaded from http://genome.ucsc.edu/cgi-bin/hgTables. I use this as refFlat.

head of gtf:

1    hg19_refFlat    exon    11874    12227    0.000000    +    .    gene_id "DDX11L1"; transcript_id "DDX11L1"; 
1    hg19_refFlat    exon    12613    12721    0.000000    +    .    gene_id "DDX11L1"; transcript_id "DDX11L1"; 
1    hg19_refFlat    exon    13221    14409    0.000000    +    .    gene_id "DDX11L1"; transcript_id "DDX11L1";

I think the problem is this gtf file. Does anyone know how to get the correct refFlat file of human?

The error from CollectRnaSeqMetrics is:

Exception in thread "main" net.sf.picard.annotation.AnnotationException: Wrong number of fields in refFlat file /home/JCheng/UCSC_RefSeqGenes_GRCh37_hg19_withoutChr.gtf at line 1
    at net.sf.picard.annotation.RefFlatReader.load(RefFlatReader.java:80)
    at net.sf.picard.annotation.RefFlatReader.load(RefFlatReader.java:66)
    at net.sf.picard.annotation.GeneAnnotationReader.loadRefFlat(GeneAnnotationReader.java:37)
    at net.sf.picard.analysis.CollectRnaSeqMetrics.setup(CollectRnaSeqMetrics.java:96)
    at net.sf.picard.analysis.SinglePassSamProgram.makeItSo(SinglePassSamProgram.java:102)
    at net.sf.picard.analysis.SinglePassSamProgram.doWork(SinglePassSamProgram.java:55)
    at net.sf.picard.cmdline.CommandLineProgram.instanceMain(CommandLineProgram.java:177)
    at net.sf.picard.cmdline.CommandLineProgram.instanceMainWithExit(CommandLineProgram.java:119)
    at net.sf.picard.analysis.CollectRnaSeqMetrics.main(CollectRnaSeqMetrics.java:88)

[Biocode_user]

Hello,

I am getting same error and could anyone please help me. I used gtfToGenePred to convert my gtf file to refFlat file. and my refFlat looks like

Ec-00_000010.1  chr_00  -       149     6731    149     6731    10      149,897,1535,2091,2535,3474,4006,4702,6245,6709,        428,1100,1674,2268,3070,3557,4155,4968,6363,6731,
Ec-00_000020.1  chr_00  -       28572   29122   28572   29122   2       28572,28937,    28582,29122,
Ec-00_000030.1  chr_00  +       29412   32214   29412   32214   1       29412,  32214,
Ec-00_000040.1  chr_00  +       34287   34360   34287   34360   1       34287,  34360,
Ec-00_000050.1  chr_00  -       36705   39329   36705   37902   3       36705,37422,39143,      36870,37944,39329,
Ec-00_000060.1  chr_00  +       43007   44099   43007   44099   3       43007,43404,43829,      43046,43455,44099,

and when I try to run

java -jar /software/shared/apps/x86_64/picard-tools/1.56/CollectRnaSeqMetrics.jar \
    REF_FLAT=chr_00.refFlat.txt \
    RIBOSOMAL_INTERVALS=null \
    STRAND_SPECIFICITY= NONE \
    CHART_OUTPUT=b2_cOLLECTrna.pdf \
    METRIC_ACCUMULATION_LEVEL=ALL_READS \
    INPUT=out.prefix.bam \
    OUTPUT=B2_CollectRNAMetrices

I end up getting

Exception in thread "main" net.sf.picard.annotation.AnnotationException: Wrong number of fields in refFlat file chr_00.refFlat.txt at line 1
        at net.sf.picard.annotation.RefFlatReader.load(RefFlatReader.java:80)
        at net.sf.picard.annotation.RefFlatReader.load(RefFlatReader.java:66)
        at net.sf.picard.annotation.GeneAnnotationReader.loadRefFlat(GeneAnnotationReader.java:37)
        at net.sf.picard.analysis.CollectRnaSeqMetrics.setup(CollectRnaSeqMetrics.java:137)
        at net.sf.picard.analysis.SinglePassSamProgram.makeItSo(SinglePassSamProgram.java:101)
        at net.sf.picard.analysis.SinglePassSamProgram.doWork(SinglePassSamProgram.java:54)
        at net.sf.picard.cmdline.CommandLineProgram.instanceMain(CommandLineProgram.java:175)
        at net.sf.picard.cmdline.CommandLineProgram.instanceMainWithExit(CommandLineProgram.java:118)
        at net.sf.picard.analysis.CollectRnaSeqMetrics.main(CollectRnaSeqMetrics.java:101

Could anyone help me figure out what wrong with Refflat file?

[opplatek]

You are missing the gene name at the first column. Check solinvicta comment to fix this. Basically you have to run gtfToGenePred -genePredExt and then move the gene name (column 12) to the first column.

Answer 1

[Pierre Lindenbaum]

from the sources refflat should have the following columns:

 enum RefFlatColumns{GENE_NAME, TRANSCRIPT_NAME, CHROMOSOME, STRAND, TX_START, TX_END, CDS_START, CDS_END, EXON_COUNT, EXON_STARTS, EXON_ENDS}

You'd better use: http://hgdownload.cse.ucsc.edu/goldenPath/hg19/database/refFlat.txt.gz

[juncheng]

Thanks you so much!

[juncheng]

Actually, does "chr1" need to be changed to "1"? My reference is "1" form.

[Pierre Lindenbaum]

yes, very probably

[ATCG]

is there a refFlat.txt for ENSEMBL? Thanks!

Answer 2

[solinvicta]

Found this link and the process seemed to work - pretty much you output a few extra fields with different options in the original command and then trim them back:

https://www.snip2code.com/Snippet/77082/Convert-gene-annotations-from-GTF-to-gen