GTF based salmon index

[parvathisudha]

Hi Salmon team,

Thanks for the tool! I am trying to create the salmon index for GRCh38 using Gencode. When I did the quantification, even though I added GTF file, "quant.genes.sf is only showing the Transcript id's not gene id's. Can you please tell me how to solve this issue? Is there a way to create a GTF based salmon index file for GRCh38?

Thanks Parvtahi.

[yeodynasty]

I have a similar problem. Attached are:

1. gtf file, where clearly, the gene_ id and transcript_id are provided
2. quant files are as followed for gene and transcripts

3. my command as followed:

/gpfsdata/apps/salmon-latest_linux_x86_64/bin/salmon quant \ -i /gpfshome/hockchuan/SALMON/GCF_900626175.2_cs10_index \ -l ISR \ -1 /gpfsdata/JangiLab/hockchuan/170302/2.Trimmomatic_output/clean_HEADBANDSTEM_1.fastq.gz \ -2 /gpfsdata/JangiLab/hockchuan/170302/2.Trimmomatic_output/clean_HEADBANDSTEM_2.fastq.gz \ --seqBias \ --gcBias \ --posBias \ --incompatPrior 0.0 \ --geneMap /gpfsdata/JangiLab/hockchuan/cs10_reference_genome/GCF_900626175.2_cs10_genomic.gtf \ --recoverOrphans \ --allowDovetail \ --threads $NSLOTS \ --dumpEq \ --minScoreFraction 0.65 \ --writeMappings /gpfshome/hockchuan/SALMON/MAP/HEADBANDSTEM \ --fldMean 250 \ --fldSD 25 \ --writeOrphanLinks \ --writeUnmappedNames \ --quiet \ -o /gpfshome/hockchuan/SALMON/HEADBANDSTEM_quant

fewLines.gtf.txt quant.genes.txt quant.txt

[yeodynasty]

Any idea what went wrong?

[rob-p]

Is there any output to the terminal when salmon is running that would suggest it couldn't interpret the GTF properly? Can you share the salmon log file?

[parvathisudha]

Any idea what went wrong?

I had that issue while using Salmon version 0.9.1. Once I upgraded the salmon version, I could index the genome properly and I obtained gene-level TPM for the samples.

[yeodynasty]

This is the initial output log, where it reports an inccorrect gene annotation:

---

Version Info: This is the most recent version of salmon.

| Loading contig table | Time = 13.512 s

size = 16145665

| Loading contig offsets | Time = 382.03 ms

---

| Loading reference lengths | Time = 9.4861 ms

---

| Loading mphf table | Time = 2.4236 s

size = 1057188904 Number of ones: 16145664 Number of ones per inventory item: 512 Inventory entries filled: 31535

| Loading contig boundaries | Time = 4.031 s

size = 1057188904

| Loading sequence | Time = 1.983 s

size = 572818984

| Loading positions | Time = 14.658 s

size = 942318702

| Loading reference sequence | Time = 1.4932 s

---

| Loading reference accumulative lengths | Time = 10.959 ms

Error: invalid feature coordinates (end<start!) at line: NC_029855.1 RefSeq gene 406748 107842 . + . gene_id "A5N79_gp28"; db_xref "GeneID:27215502"; exception "trans-splicing"; gbkey "Gene"; gene "nad2"; gene_biotype "protein_coding"; locus_tag "A5N79_gp28";

---

After I remove the erroneous entry, there is no more complaint:

---

Version Info: This is the most recent version of salmon.

| Loading contig table | Time = 14.648 s

size = 16145665

| Loading contig offsets | Time = 336.77 ms

---

| Loading reference lengths | Time = 10.195 ms

---

| Loading mphf table | Time = 2.3113 s

size = 1057188904 Number of ones: 16145664 Number of ones per inventory item: 512 Inventory entries filled: 31535

| Loading contig boundaries | Time = 4.881 s

size = 1057188904

| Loading sequence | Time = 1.7554 s

size = 572818984

| Loading positions | Time = 13.626 s

size = 942318702

| Loading reference sequence | Time = 1.5082 s

---

| Loading reference accumulative lengths | Time = 12.272 ms

---

However, the *.sf files are the same as previous ones, i.e. no gene level results.

[yeodynasty]

Interestingly, the same gtf file can be used to obtain gene level counts using HTSeq-count in OmicsBox, suggesting that the file is working fine.

[yeodynasty]

I have used tximport to complete the job instead.

[rob-p]

Thanks for letting us know. Tximport is the recommended way to accomplish this.