MEGA Analysis Issue on High-Performance Computer
1
0
Entering edit mode
4 weeks ago
Giorgio • 0

Dear MEGA Community,

I am encountering an issue with the MEGA software during phylogenetic analysis on my powerful computer. Despite having 16 GB of RAM, an Intel Core i9 13900H processor, and a GeForce RTX 4060 graphics card, the analysis stops working after about 1 hour and 30 minutes and does not complete, it only complete 37 of 50 boostraps.

I have ensured that I am using the latest version of MEGA and that other applications are not consuming excessive resources during the analysis. I have also checked that there is sufficient disk space available.

Could you please advise me on how to troubleshoot this issue or provide any insights into why this might be happening?

Thank you for your assistance.

Best regards, Giorgio.

phylogenetic-analysis MEGA • 664 views
ADD COMMENT
2
Entering edit mode
4 weeks ago
Mensur Dlakic ★ 27k

I have ensured that I am using the latest version of MEGA and that other applications are not consuming excessive resources during the analysis. I have also checked that there is sufficient disk space available.

You have answered your own question, probably without realizing. If any kind of analysis couldn't finish (what is the error message?), it has to be a lack of resources (memory, disk space) or a lack of time (having a slow computer or didn't wait long enough).

Your computer is not slow, so that shouldn't be a problem. We still don't know from your description whether there was an error of some kind, or the process simply slowed down enough that you thought it stopped.

I will take your word that disk space is sufficient. That leaves memory. Contrary to your opinion, 16 GB does not qualify as "high-performance" or "powerful" even though you have a fast and modern computer. That may be enough to run several applications for personal use, but not necessarily for phylogenetic reconstructions, especially if you have a large alignment. That's another thing you didn't tell us that matters more than your computer specs: how large is the alignment? Finally, having a GPU doesn't necessarily mean it is utilized by any given program. If MEGA is actually using your GeForce RTX 4060, then it is even more likely it ran out of memory, because your GPU's 8 GB is really not that much when it comes to phylogenetic analyses.

ADD COMMENT
0
Entering edit mode

16GB of RAM is nothing. My phone has 12GB. My server has 1TB, and even that isn't always enough

ADD REPLY
0
Entering edit mode

The error that occurs is that the program freezes and displays "not responding" before closing. What's suspicious is that I run the same program, with the same parameters and alignment, on another computer, which is inferior to mine (8 GB of RAM and an Intel Core i5), and it is able to obtain the phylogenetic tree, albeit taking considerably longer, approximately 3 days. The alignment consists of 270 sequences. Why can the computer with lower performance than mine obtain the result and mine cannot? Could it be a problem with my Windows?

Thank you very much for your help.

ADD REPLY
0
Entering edit mode

It is a safe bet that your problems are memory-caused. It could be that the other computer had a better caching system that allows it to work with disk space rather than RAM. The price for that is that it takes much longer, which tracks with 3 days. And yes, it could be a Windows memory management problem.

The alignment is like a matrix: giving one dimension is not sufficient. I have an alignment that is 150 sequences long, so it may be tempting to conclude it is smaller than yours. Then I tell you it is ~15,000 positions wide, and that's something different altogether. This may give you a sense of scale: an alignment that is about 150 x 2500 can fit into a 8 GB GPU, but not 4 GB.

Many people swear by MEGA, presumably because of its GUI and because it automatically writes figure legends. It has been a while since I used it so maybe my opinion is outdated, but back then there were many better (and faster) choice with more options and without a GUI memory overhead. You may want to consider modern stand-alone ML programs such as RAxML and IQ-TREE.

As has been pointed out and also you might know, many smart phones have 64+ GB. I suggest you re-calibrate your opinion that a computer with 16 GB is "powerful" or "high-performance" when it comes to bioinformatics applications. Back in 2016 I bought two desktops with 64 GB of RAM each that are used today for routine projects that have lower memory requirements. The other 4 servers with 16 GB RAM are either retired or used for class projects. I bought two 256 GB servers in 2020 that will soon need an upgrade. The only reason I can still get by is because of 1.5 TB nodes on a local cluster.

ADD REPLY

Login before adding your answer.

Traffic: 1644 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