I was wondering if this laptop would be sufficient for small dataset/sequence analysis? My focus currently is mainly around cancer genomic datasets. I hope to analyze RNA-seq data + genome & transcriptome data from TCGA.
Tasks/Purposes
(Primary) Quantifying gene expression and performing differential expression analysis (likely using EdgeR/DeSEQ2)
Visualization, likely to be done on web tools however (e.g UCSC Xena)
Identification of somatic variations (mutations)
I may do a little programming/code development, but it will not be intensive
This will not be my primary laptop, and I don't intend to lug it around much, if at all. I am comfortable in the event it takes multiple days to process/run.
Specs
Processor: 11th Gen Intel i7-1165G7 (2.80 GHz, up to 4.70 GHz turbo, Quad Core)
RAM: 16 GB DDR4 3200MHz
Storage: 1 TB PCIe SSD
I may purchase a 3TB external drive, and maybe 1TB cloud storage via AWS
Graphics: Integrated Intel Iris Xe Graphics
As for the OS, I'm looking into doing Windows 10 + Linux/Ubuntu through WSL2. It is a Thinkpad.
Other:
Not sure if it's relevant, but I already have a large external Samsung monitor, and external keyboard + mouse
My computer needs may not be best matched with yours, as I have 3 desktops with 16 Gb that are useless for any of my real applications. Yet another one at 32 Gb is being relegated to small tasks, so for me the configuration that is useful on a daily basis starts at 64 Gb RAM. The last two desktops I bought were at 256 Gb RAM each, and I already wish I had at least one at 512 Gb.
I think you got lots of useful advice here, so I suggest you read everything. Based on what Gordon Smyth wrote, it seems like 32 Gb would work nicely for your intended work.
I personally use a Windows 10 laptop similar to one you are considering except that I have 32MB RAM and 2TB SSD. I find it perfectly satisfactory for all the analyses you have described, including short read alignment with Rsubread. I do not bother with WSL2. Instead I run gcc compilers etc as needed via the standard R developer pack for Windows. Windows supports a large number of programming productivity tools such as tortoisegit or Beyond Compare. I do use external Linux servers for running memory hogs like 10x CellRanger or for unix-only tools like Salmon. Large-scale integration of single cell data from 100s of thousands of cells also needs a bigger machine, but you make it clear that is not your purpose. If you want to use lots of Unix tools, then a non-Windows machine might be better for you. But for running R and Bioconductor with TCGA, a highly configured Windows laptop is a superb tool in my experience.
I agree that 16 Gb RAM is probably not enough. Buy as much RAM as you can afford, even at the expense of disk space. Disk is cheap and easy to add later, compared to RAM.
If you don't plan to "lug it around much", think long and hard about whether you really want a laptop. A laptop is always more expensive than a desktop PC with the same hardware configuration. Adding hardware or doing repairs is also more expensive, and sometimes almost impossible depending on what the laptop needs. Laptops have always been a compromise, in which ergonomics, expandability and cost are traded for mobility. External keyboards are more ergonomic than laptop keyboards, and are more easily replaceable. Laptop monitors are smaller than their desktop counterparts. As well, many recent laptops don't have Ethernet ports. You can buy a USB/Ethernet adaptor, but that adds to the cost. You want hard-wired network for bioinformatics, because WiFi ranges from superb to appalling, sometimes depending on things you can't control.
Give serious consideration to going 100% Linux. These days, the barriers to using Linux are almost entirely psychological. My own experience has been that when I am forced to work on in Windows, I find myself cursing and thinking "this would be so much more straightforward on Linux". Despite the hype, WSL2 is still a work in progress. Although you can get full command line Linux with in a few steps, getting the full graphical environment, as of this writing, is still more complicated. Once you have a complete WSL2 environment, you still have all the steps that you would do anyway on a dedicated machine. These would include installing whatever doesn't come by default on Ubuntu, setting system parameters, managing security, setting up devices. If you've ever managed a virtual machine, you would have an idea of all the things that entails. Also keep in mind that in addition to updating Windows 10, you will also need to keep updating Ubuntu. In essence, WSL2 gives you all the work of managing 2 computers.
One final thing with regard to Windows - There is an ongoing discussion about Windows 11 and its requirements for hardware security features such as TPM and UEFI. Google these terms and be prepared for hours of discussion on the potential nightmares Windows 11 may introduce. The worst case is that even some computers being sold today will not support TPM in hardware, and that even recent computers will not be upgradable to W11. My take is that the latter will have workarounds, but I would advise anyone thinking of buying a W10 machine today to read up on this debate.
To summarize, if you go the Windows10/WSL2 route, you are needlessly adding a layer of complexity and ongoing system administration headaches. A pure Linux machine has the advantage of simplicity.
I would suggest to ramp up the ram and also the storage, since 16GB is not enough for serious bioinformatics work. I would also like to suggest you to go fully Linux oriented since almost all the programs are available for Linux and the so existed gap between windows and Linux is shrunken so much. If you plan on keeping the system for a while I would suggest to go for a custom PC build (desktop) for the purpose with ,ore ram, storage and also cooling.
How "small" are these datasets?