Sufficient Specs for a laptop to work in bioinformatics
1
0
Entering edit mode
2.3 years ago

Hello everyone,

I am planning to buy a new laptop to work in bioinformatics, I have intermediate knowledge in bioinformatics. Do we have to buy a mac to work in bioinformatics or is it okay to work on other laptops with good specifications. What kind of specs are usually required to work in bioinformatics efficiently? Thanks in advance. Looking forward to your suggestions!

hardware NGS • 5.6k views
ADD COMMENT
3
Entering edit mode
2.3 years ago

Depends on what sort of bioinformatics you will be doing and what sort of other resources you will have access to.

I have a very ordinary laptop (i5, 8GB RAM), but I do next to no hardcore bioinformatics on it. I have a 10 year old desktop with a Xeon E3 and 32GB RAM for running interactive R, jupyter and IGV sessions.

For eukaryotic sequencing type things (Definitely variant analysis or de-novo assembly, but also most likely RNAseq, ChIP-seq etc), few laptops are going to cut it either way, and you need access to a server - either your own server, or on a cluster, or in the cloud. If you are doing prokaryotic sequencing stuff, then you can probably get away with mostly using a local machine. The same probably applies if you are doing methods development, or lightweight systems biology simulations (the biggest simulations go all the way up to needing to run on national-level supercomputers).

With all that said, I tend to buy people in my lab a windows i7 desktop machine with 32GB of RAM and 1TB disk. This allows them to do most interactive things they will ever need to do. For non-interactive we use the institutional cluster. We also store all data on network locations, and local disk is only used for installing software and temporary files. Very few laptops will ever provide enough disk space to store full, raw, datasets from eukaryotic genomics.

With the advent of WSL on Windows, it makes very little difference whether you choose windows, mac or linux - pick which ever you like using the most, or fits your budget best. Mac is probably currently the fastest (with its M1 line of apple silicon hardware), but its way more expensive. Faster CPUs will make your life easier, but more RAM will make the difference between being able to do something and not being able to do it. Given the choice, choose more RAM over a faster CPU - I wouldn't buy anything with less than 16GB if its going to be your main system, and really you want to be looking for 32GB of RAM.

ADD COMMENT
0
Entering edit mode

I believe that when it comes to computing capacity/dollar the Mac M1 is the best deal by far. The Mac M1 CPU will outperform any other comparable CPU by usually double if not triple in speed of execution. The only downside may be that RAM can only be expanded to 16GB, but I expect that to change with the new processors that are coming out.

https://browser.geekbench.com/search?q=M1

vs

https://browser.geekbench.com/search?q=i7

ADD REPLY
1
Entering edit mode

Depends if you really care about speed or not. In general, if I'm doing it on my laptop, speed is a secondary concern to the raw "will it run or not?", which is usually determined by RAM.

ADD REPLY
1
Entering edit mode

Its a bit of a Sam Vime's boots sort of a situation. Its all very well and good saying that a $3000 machine gives you more flops per dollar, but if you've only got $1500, that's kind of meaningless.

ADD REPLY
0
Entering edit mode

Here is where the logic breaks down. The M1 processor is both faster and cheaper even at "low" end.

A Macbook Air with 16GB of ram costs $1080 and as such, it is cheaper than any comparable laptop or even a desktop with not even half as good performance

As an actual observed use case: when I map 100 million reads in 8 threaded mapping my MacBook Air is just twice as slow than when I run the same job as 24 threads on a server that costs $30K.

Of course, the server has tons of memory and will fit jobs that the Macbook Air won't. But if we're talking running RNA-Seq with salmon over the say mouse genome ... I think the MacBook Air will run circles around the competition (I will test this out soon and report back)

ADD REPLY
0
Entering edit mode

My 40 core server only cost $10k, but we probably get a good deal on our institutional agreement :)

I wonder if you've hit the point of diminishing returns with 24 threads? I wonder how it would do aligning three samples with 8 threads each or 6 samples with 4 threads each (as 6 samples is probably the most common size of an RNAseq experiment).

ADD REPLY
0
Entering edit mode

Unfortunately, I don't think I have the budget to get a 32 GB ram so probably 8 or 16 could be the highest I can get for now. Thanks a lot for your help, it has eased the way for buying a laptop now.

ADD REPLY
0
Entering edit mode

Hello... Did you buy the laptop with 16 gb?

ADD REPLY

Login before adding your answer.

Traffic: 2929 users visited in the last hour
Help About
FAQ
Access RSS
API
Stats

Use of this site constitutes acceptance of our User Agreement and Privacy Policy.

Powered by the version 2.3.6