I suppose you are joking when you ask about the ubuntu configuration for NGS analysis. This was the best joke I have seen today.

Thanks.

Hi Alireza,

You can predict (imprecisely) the required space for any sequencing, but be aware that sequence technology, instrument and output format affect the space you need:

For your 1, 2 and 3 points.

1) Calculate your sequencing output in Gbp with the formula:

output = (Genome or panel size * coverage) * N° of samples

2) Calculate the number of lanes/flow cells you're gonna use by checking the producer specification for each instrument:

lanes required = output / expected Gbp per flow cell

3) You can use this illumina guide to convert flow cells to disk occupancy depending on the illumina instrument you use: https://emea.support.illumina.com/bulletins/2018/01/approximate-sizes-of-sequencing-run-output-folders.html

Es. for case 2, let's take 50Mb as average exon size (it's an example)

50 Mb * 120 * 5 = 30.000 Mb = 30 Gb, this is the desired output

Let's say you sequence in 2x150bp and use the NextSeq 550 with a mid output kit, for an output of abut 30 Gb. From the table you can see that that this run would occupy approximately 62 - 66 GB.

These are just the FASTQs, for the BAM I usually calculate about the same size as the fastq if they are sorted and if you keep only the mapped reads (I never worked on human though), but this really depends on your genome, aligner and settings.

I don't know if PacBio offers a similar table, but the size really depends on wether you're using long reads or HiFi reads (they require a smaller coverage and so less data). For Long reads I think we are in the order of 3-4 TB, for HiFi you should stay below 500GB-1TB, including the assembly.

At first glance I think you'll be ok with at least 5 TB, but I'd go with 8 TB at minimu (better more than less space).

You can run your analysis on an external HDD but it is not a good idea because the reading and driving speed will be limited and HDDs are not very safe for working with data (they're not very safe in general).

Also, be careful with the computing power you have (CPUs and memory) if you're not using a cluster. Alignining WGS data and in particular PacBio is a really demanding task, PacBio may require thousands of CPU hours.

Umberto Palatini Many thanks for your comprehensive reply. I will consider that and go for my data analysis. 🙏

Abhijeet Singh Learn not to mock other people. Maybe I just used wrong words for my question.