This is a test version of Biostars. For the public version, visit https://www.biostars.org.
Average Insert Size When Using Transcript Quantification Tools Such as Salmon?

I'm currently in the process of uploading my dataset to the gene omnibus (GEO), but the meta data form asks for an average insert size, but I'm not sure where to find this information. I used the transcript quantification tool Salmon in my analysis and I don't know if this information even exists given I didn't map the reads against a reference genome

Question: Is it possible to get an average insert size when mapping against a transcriptome using the newer transcript quantification tools such as Salmon

Any help would be greatly appreciated

rna-seq salmon next-gen rna-seq
• 3,853 views • link
updated by Rob Rob • written by Dru Zod • Dru Zod

This doesn't actually the question, but the average insert size field is optional, so you can leave it blank if you don't know it.

• 1 views • link
written by igor igor

that's good to know. thanks

• 1 views • link
written by Dru Zod • Dru Zod

2 answers

Is the data paired-end? In this case, salmon estimated the fragment length distribution. In the quantification directory the file libParams/flenDist.txt contains the parameters of the normalized distribution of observed fragment lengths. Each of the 1001 numbers gives the probability of observing a fragment of the same size. You can then extract any information you wish (mean, median, standard deviation, etc.) from this distribution.

• 1 views • link
written by Rob Rob

Yes, my data was paired-end. I don't think I have this file saved, but I can re-run the analysis on the same files and take a look and mark this as the correct answer if it all works. Thanks

• 1 views • link
written by Dru Zod • Dru Zod

Thank you - both for your great work with Salmon and for following up questions like this!!

• 1 views • link
written by cursons.j • cursons.j

It's highly likely that whoever did the library preparation will have this information. It's quite likely that they ran the samples on a bioanalyzer, or tape station, or something similar and would have these values from it. So ask them.

Alternatively, align a subset to the transcriptome and get the values from that.

• 1 views • link
written by Devon Ryan Devon Ryan

If the wet-lab data are unavailable, you could also align a subset of the reads and get the insert size based on the paired-end information. picard CollectInsertSizeMetrics is an option.

• 1 views • link
written by ATpoint ATpoint

I can check with our universities core, but they did give me a file of qc information and the average insert size wasn't included, so I'm not sure if I'll have much luck

• 1 views • link
written by Dru Zod • Dru Zod

Log in to answer this question.

More posts like this