Memory and time Estimates for 100x genome alignments

Question

[Nujra]

Hi, I am generating sets of data up to 100x human WGS. Our lab is buying a server so I was wondering the normal time and memory this would require with say 10-15 threads using either bwa/bowtie/minimap. Thanks

[ANILA]

Hi, please convert the following codes into r codes. or help me if anyone knows how to perform bagging with support vector machine classifier in r.

from sklearn.svm import LinearSVC
from sklearn.ensemble import BaggingClassifier
import hasy_tools  # pip install hasy_tools

# Load and preprocess data
data = hasy_tools.load_data()
X = data['x_train']
X = hasy_tools.preprocess(X)
X = X.reshape(len(X), -1)
y = data['y_train']

# Reduce dataset
dataset_size = 100
X = X[:dataset_size]
y = y[:dataset_size]

# Define model
svm = LinearSVC(random_state=42)
model = BaggingClassifier(base_estimator=svm, n_estimators=31, random_state=314)

# Fit
model.fit(X, y)

[GenoMax]

What does this have to do with the original question? Please create a new thread to ask a new question.

Answer 1

[Chris Dean]

Have you considered generating synthetic reads from the human reference genome with wgsim or art_illumina to answer this question?

You can use these tools to simulate this scenario at the desired read depth, run the data through each aligner, and record the amount of time to completion using the /bin/time command (see Linux manual page).

Of course this depends on having access to the necessary hardware requirements (e.g., disk space and RAM).

I hope this can help you get a rough idea about the run time!

[Nujra]

I have preexisting data. The issue is that my nodes are limited in memory and I haven't been able to successfully complete the alignment. I was therefore wondering whether someone with personal experience with this type of data has some sort of insight on these parameters.