Bioinformatics Chat

I need a book that explains Next-Generation Sequencing. If someone has one, please send the book.

If I were to start in 2026, what would you advise me to do and how should I begin?

Python Linux

Alright... lets get this started !

I'm new here what well you recommend me should i learn bioinformatics or not. I am BIOTECHNOLOGY student.

Use line commands, learn Linux, read papers, and just practice until you can teach it

How should I do that?

Do you want a road map?

Yes

Please I could use a roadmap as well

use AI for everything, you're welcome

Hi I would like to ask for biological interpretation. WGCNA was performed on quantile-normalized RNA-seq data from a mouse melanoma model. Hub genes was selected according to MM and GS for immunotherapy response. After that, LASSO was applied to narrow down the number of candidates. 4 genes were obtained. However, in a human clinical cohort (Gide et al. 2019), using these 4 genes in a glm model to predict clinical response, some of these genes reverse their coefficient signs. The AUC is pretty great around 0.7.

Hi! It's seems to that results are conterdirectory in a practice although to high accuracy. I think that the reason for this contradiction is due to interspecies differences. Such a situation is not uncommon in translational studies, especially when moving from a mice model to a human trial. The biology of two species has a lot of distinctive features. However, AUC ~ 0.7 is a good result for such a complex clinical endpoint. I'd like to recommend you to use a model as a prognostic signature but not as a causal set of genes

I've summarized the main steps # samples combo <- c("Combo_E28", "Combo_E29", "Combo_E30", "Combo_E31", "Combo_E32") X_combo <- as.matrix(gene_expression_clon[combo, _hubs]) Y_combo <- c(0, 0, 1, 0, 1) names(Y_combo) <- combo library(glmnet) set.seed(123) lasso <- glmnet(X_combo, Y_combo, family = "binomial", alpha = 1) model <- glm(Respuesta ~ geneA + geneB + geneC + geneD , data = df_model, family = "binomial") weight_model <- summary(model)$coefficients values <- weight_model[genes, "Estimate"] df_signature <- data.frame( Gene = genes, Weight = as.numeric(values), stringsAsFactors = FALSE )

Hello! I'm brazilian. I'm doing a specialization in Bioinformatics, and I want to apply for a Master's program, but I haven't decided on my research area yet. How did you get interested in your current research area? Can you tell me something about it, please?

Hi everyone, I am analyzing an RNA-Seq dataset consisting of 46 samples encompassing 3 distinct biological clones treated with two drugs drug A and B and their respective two control: control A and B My primary goal is to isolate genes exhibiting significant differential co-expression network connectivity between Treatment A and the Control, while strictly controlling for baseline clonal variation. To achieve this, I designed a hybrid workflow combining WGCNA (Weighted Gene Co-expression Network Analysis) and PRANA (Pseudo-value Regression Analysis for Network Association). I would highly appreciate your feedback on the validity of my downstream network isolation steps. sub_TOM <- TOM_clon[top10, top10] cyt <- exportNetworkToCytoscape( adjMat = sub_TOM, weighted = TRUE, threshold = 0.02, nodeNames = top10 ) edges_conjunta <- cyt$edgeData clone_matrix <- model.matrix(~ 0 + as.factor(meta_combo$clone)) colnames(clone_matrix) <- paste0("Clone_", levels(as.factor(meta_combo$clone))) idx_ctr <- which(meta_combo$class == "Ctr") idx_combo <- which(meta_combo$class == "Combo") exp_ctr <- rnaseqdat_filtrado[idx_ctr, top10] exp_combo <- rnaseqdat_filtrado[idx_combo, top10] clones_ctr <- clone_matrix[idx_ctr, -1, drop = FALSE] clones_combo <- clone_matrix[idx_combo, -1, drop = FALSE] pcor_ctr_full <- pcor(cbind(exp_ctr, clones_ctr))$estimate pcor_combo_full <- pcor(cbind(exp_combo, clones_combo))$estimate pcor_ctr <- pcor_ctr_full[top10, top10] pcor_combo <- pcor_combo_full[top10, top10] edges_conjunta$pCor_en_CTR <- apply(edges_conjunta, 1, function(row) pcor_ctr[row["fromNode"], row["toNode"]]) edges_conjunta$pCor_en_COMBO <- apply(edges_conjunta, 1, function(row) pcor_combo[row["fromNode"], row["toNode"]]) edges_conjunta$Delta_pCor <- edges_conjunta$pCor_en_CTR - edges_conjunta$pCor_en_COMBO write.table(edges_conjunta, file = "WGCNA_PRANA_Edges_PartialCor.txt", sep = "\t", row.names = FALSE, quote = FALSE) write.table(cyt$nodeD

Hii Matheus I will start my master this October, it will be in medical sciences but part of my research will need bioinformatics skills Just find something you are really interested about and than look for a lab that do the same type of research Don't need to worry to much about it, if you like a field you could try to make a project that answer today problems in that area Bioinformatics is very useful, you could do research about gene expression, proteomics, metabolism, new candidate drugs using those new computational and IA tools, etc Boa sorte mano, ti desejo sucessos😊

Thank you so much! Good luck to you in your master! Vlw mano, obrigado mesmo!

Hello! It’s a pleasure to meet you, and welcome to this fascinating field like Bioinformatics. When I was starting my Master's, I chose to dive into the study of neurodegenerative diseases. For me, the motivation was simple yet powerful: the brain is perhaps the most complex system we know, and using computational tools to understand how it fails in disease felt like doing cutting-edge detective work. Bioinformatics gives us the unique power to look at molecular puzzles and find hidden links that can eventually help people. So that’s attractive and tricky. My advice as you choose your path: look for the problem you find most beautiful or most challenging to solve. And.. Here’s one point: what are your career and money goals? I believe a scientist should answer this question for himself. What kind of biological puzzles have intrigued you the most during your specialization so far? Let me know in private messages, if you want, and we can look at how to shape that into a research path. Congrats, colleague 😎

Hi, I would like to ask if someone has used CWGCNA (causal inference from the WGCNA framework) ??