Forum:New Macbook M4 laptops (Pro vs Max)

Question

[tommygoe]

Hi guys,

I was looking for some advice on whether the new macbook pro M4 has sufficient power for most/all biological analyses such as RNA-Seq analysis from raw files, as well as other heavy bioinformatic tasks.

It states for macbook pro m4 -12-Core CPU 16-Core GPU 24GB Unified Memory 512GB SSD Storage¹

There is also another version of the pro having 14-Core CPU 20-Core GPU 24GB Unified Memory 1TB SSD Storage¹

While the Max has 14-Core CPU 32-Core GPU 36GB Unified Memory 1TB SSD Storage¹

Which of these three is more than sufficient for the most demanding bioinformatic tasks?

[andres.firrincieli]

Define most demanding because for what I usually do (metagenomics) none of the above would be sufficient.

[tommygoe]

Full RNA SEQ analysis from raw reads for both short and long read sequencing

Answer 1

[Mensur Dlakic]

You already got the answer multiple times, but somehow can't seem to accept it. Laptops are not meant for the type of job you want to do, or for any "heavy bioinformatics" jobs.

That goes both for the RAM values your cited and for storage. Let's go with the largest numbers you quoted: 36GB Unified Memory and 1TB SSD storage. Those are seriously inadequate for RNAseq applications, where the RAM needs to be in hundreds of GB and the storage in tens of TB.

You can do all the regular jobs with these laptops, plus serious graphics applications. You can do machine learning, model fairly large proteins and a bunch of other applications that do not require large storage space but benefit from a generous GPU memory. But no, laptops are not meant for aligning big reads datasets to genomes. Even if you had 5x more RAM and 10x more disk space, the laptop cooling system would be seriously stressed working non-stop for many hours. That is a recipe for shortening the life of a laptop.

[dsull]

It depends on how "heavy" we're talking; if OP is a student just looking to analyze six or so RNAseq samples for an undergrad thesis project, a laptop is fine. If OP works in a lab where a billion reads is coming off the sequencing machine every week (with an occasional novaseq being performed), then no.

Edit: By "heavy", I had assumed the OP was referring to the latter but if it's just a one-time analysis of 20 samples, I think a laptop is fine. Not sure why one would buy an expensive laptop just to analyze 20 samples though...

[Mensur Dlakic]

Any question can be answered half a dozen ways if we add enough qualifications. If a person has a month to spare and only a handful of samples, even 16 GB of RAM and 1 TB SSD could do the trick. If they are willing to offload the data after each analysis onto a secondary drive, even 512 GB SSD might be enough. Then again, if they are mapping to a 20-50 Gbp genome, the memory and disk storage numbers we are talking about here would be laughable.

The question was about using this particular laptop for RNA-Seq analysis from raw files, as well as other heavy bioinformatic tasks. To that question the answer is no, this type of a computer is not a good choice. No commercial laptop is, for that matter.

[dsull]

that's fair, agreed

Answer 2

[Shred]

None of them. Most of the bioinformatics tasks requires memory rather than efficient cores (like the M series from Apple). Aligning RNA-seq reads to a reference genome requires lots of RAM (usually around 50-60GB). Pay attention also to the ARM architecture which may not be compatible with some libraries written for x64 architectures.

[dsull]

…you do not need 60 gb of RAM for RNAseq.

(I do agree about prioritizing buying something with more memory though).

[tommygoe]

Hi, Thanks a lot for your reply, much appreciated. Are there any laptops that can do this or is it a case of using a server to do this (then assume laptop doesn’t matter too much)?

[ATpoint]

Many good laptops can align a small dataset if you give it some time, the question is the scale. Do you have like dozens of samples coming in per month, then consider a solid setup like a server, but if your project has like a nw dataset coming every 6 months you can stick to a laptop. It all really comes down to the amount of data you need to process.

Generally, laptops are meant to be convenient and portable at the price of reduced heat management compared to towers or servers, so they're not meant or designed for heavy duty. Keep that in mind.

[tommygoe]

Hi, it would be around 20 samples, would M4 pro be sufficient for doing this without a server/google colab?

[dsull]

M4 pro can handle processing 20 samples; that's not exactly considered "heavy".

(Unless you're doing a heavy task with those 20 samples like making assemblies from them. But for standard RNAseq analysis, you'll be fine.)