Nominal length vs insert size

Question

[nusrat.bot]

Hello Everybody,

Can I use nominal length in sra data (paired) as insert size??

Answer 1

[Brian Bushnell]

Not if the insert size has an impact on your analysis. Insert size is unrelated to read length for Illumina data. You can calculate the insert size distribution by mapping or merging the pairs (e.g. BBMerge).

Additionally, insert size estimates from lab analysis are often very different from those produced by bioinformatic analysis. For bioinformatics, the latter is necessary.

[nusrat.bot]

Dear Brian Bushnell many many thanks for your reply and suggestion. I didn't use BBmerge tools before and I have little experience on the java based program. However, Some of them easy to locate the file and easy to get the result. But in case of BBmerge I couldn't locate my file from windows cmd command line. Could please please describe me details how to use BBmerge or give me a link from where can I get some tutorials or lesson

For example I am using Windows operating system. I dowloaded Java 1.8 and also BBsuite. Now in cmd shell I can locate the file of BBsuite and BBmerge but not my sequence file.

Answer 2

[shwethacm]

Yes, but use caution while using it.

Just to be sure, you can take a small subset of the reads, map them to the reference and calculate the insert size. If you use bwa mem, you can even see this in the logs. If the number seems to be close to what is mentioned on NCBI, then go ahead! It should be quick and easy!

[nusrat.bot]

Dear Shwethacm, Thank you for your suggestion. I didn't understand you very well. Please could you please make it clear for me I mean can I use any sequence as reference sequence for calculation of the insert size. Or what is bwa mem?? Is it bwa mapping??

Many thanks for your clarification.

[nusrat.bot]

When I use Galaxy interface for BWA mem of my read pairs can I use built-in index I mean which genome can I use as a reference for BWA mapping to identify the insert distance?

Many many thanks:)))

[shwethacm]

Hi Nusrat,

About references: Technically, you cannot use "any" sequence as a reference. What is the source (species) of your sequencing reads? If it is human, you can map it to the latest human assembly (hg38) to calculate insert size. In general, if you have an assembly for the species you sequenced, that could make an appropriate reference.

[nusrat.bot]

Many Many thanks shwethacm:))

[nusrat.bot]

But what if I use Pear paired-end merger tools. As because it does not require a direct insert size information.

[shwethacm]

At this point, I am confused about what your end goal is. So far, it seemed like you wanted to figure out the insert size of your PE reads, right ? You can use any alignment program to map, so long as you map it to the right reference, and have a sam/bam file in the end, you can find out the insert size.