Eight months ago I put up a thread after being hired as a computational biologist, I was scared, confused, and helpless. The original topic can be found here on an account who's password I no longer remember. I came into the field with absolutely zero knowledge in any field of bioinformatics, and only a hobbies interest in programming (thank you World of Warcraft). This thread is meant to encourage those who'd like to pursue the field, but have no bioinformatic training and those who like me have been thrown into a sink or swim scenario and are afraid of failure.
The first two months were rough. My PI has no bioinformatic background, my lab members were all purely wet lab and my bioinformatics core was largely unhelpful (more so because I didn't know what exactly I needed help with). I spent hours, and hours every single day combing through literature (much of it outdated simply because of how quickly the genomics field grows and changes), reading through Biostars postings, and Stack Overflow topics. The more time I spent trying to understand various sequencing pipelines and playing with software the easier things became.
At this point, I was coming into lab at 7 in the morning, leaving at 6 in the afternoon and 80% of my time was spent between articles and biostars. The other 20% was split between seminars / meetings and playing with programs, trying to understand their intricacies.
The third month was really when I started to feel like I wasn't a waste of time or money. Around this time my PI was asked to submit a small methods paper for a paper that a recent Graduate student had published, and I was tasked with writing up and creating all the graphs and figures for the bioinformatics portion of the paper. At this point, I had already replicated the Graduate student's paper and results a half dozen times while I was teaching myself. A few weeks of writing and a bit of R later and I had written up something that my PI was happy with, and I was given co-first author on that publication.
After the third month, I understood most of the major programs in my field, how to use them, how to apply them correctly, and just generally how to get things done. At this point is where I decided that I really started putting work into learning R, Awk and Bash scripting in order to automate processes that I was constantly having to manually repeat. I was getting tired of typing the same command line 50 times and only changing the input BAM file, and the output name.
I started with Bash, and it increased my productivity a hundred-fold. I wrote loops for just about everything I had been doing manually. What had taken me a hour before, and was often filled with all manners of human error eventually took me 10. Awk changed my life. No longer did I have to spend 30 minutes importing a text file into R, subsetting it, cbinding columns together and then writing out.
The most important lesson I learned around my fifth month was documentation and Git. I had scripts laying around everywhere, data files from who knows whom laying in who knows where. Eventually I learned how to use Git, and Markdown effectively.
Bash and Awk (and basic Unix commands) are still my most used skills to date, while I've moved most of my pipelines have since moved on to workflow languages such as Nextflow and Snakemake.
It's incredible to think back 8 months ago and read my desperate cry for help. I can now answer my own question. Yes, it is possible to learn bioinformatics on your own. Is it easy? No. But it is possible.
Looking back at some of the questions I had asked, I almost want to laugh. But then I look up and see my cubies stuffed with books, and I see my binders filled to the brim with papers and biostars topics and I remember that the reason I've learned as much as I have is because I worked hard to get here.
Everyday is a new problem, and everyday there is a new analysis I don't know how to do, or a new program I've never used before. But I'm now confident that I can accomplish what I was hired to do.
I want to seriously thank you guys at Biostars. Everyone here who takes time out of their day to answer questions, and provide invaluable amounts of information so that others can learn and grow. I can honestly say that without Biostars, I wouldn't have been able to make the switch into the field.
Props for being self-motivated enough to learn all this and for sharing your progress here! And all in 8 months!
Hey, I want to be the 2nd author ! :-)
You've done more in 8 months than I have in 2 years - congratulations! If I may ask, where do you use snakemake and nextflow? I'm yet to come across a task that I would need them for (I learn new stuff when I need to level up).
I work in a Genomics lab where I mostly process a variety of sequencing datasets, ChIP-seq, RNA-seq, MNase-seq, DNase-seq, 4sU-seq, etc.
I use
Nextflowfor these processing pipelines from trimming, mapping, to generation of bigWig tracks. Previously I used Snakemake, but because a lot of my work requires custom analysis for a specific dataset, I switched over to Nextflow as it allows me to input specific parameters for all my tools via command-line like arguments. Here's an example of one of my newer ChIP-seq pipelines (that doesn't work yet lol, i'll get back to it eventually) as an example: https://gitlab.com/c-guzman/cipher-nfI have a couple of
Snakemakeworkflows for mostly simple tasks that I'm constantly being asked to do. Analysis such as DGE or differential binding analysis of ChIP-seq data is really common amongst my lab group. Snakemake is much easier to write than Nextflow so is perfect for combining several tools and steps together very quickly. You can do this with Make, Bash, etc but I prefer Snakemake because it allows me to parallelize tasks easily and acts as a form of documentation incase someone else would like to look at the analysis.These sort of jobs come up way too often so it's nice to just be able to plug input files at the end of the day and process all my files overnight so that they're ready for downstream analysis the next morning.
I am also testing snakemake, do you have any experiences on using snakemake on HPC (moab system) ?
Unfortunately no, we have a really powerful desktop that does most of the small analysis my Snakemake files are written for.
Tell-a-tale
Being out of the field since 2010 , working as a software consultant for over 3 years and then chasing the long lost dream of coming back to the field in 2013 had its own ups and downs for me as well.
Time, persistence, patience, ambition and determination along with the help and encouragement from experts in the forum does help one and everyone in a long way to be independent in their research and also communicate science in much targeted way with clarity and am not an exception here.
I have always credited my work for Biostars and what I am today performing and helping out others will be due to the forum who helped me all along. Life is not rosy but if you stick to basics with determination it does turn in your favor be it personal or professional. The goals should be clear and ambitions should be dealt with clear vision.
I still laugh back when I look back to my questions posted 2 years back. And am happy to the place where I reached, I am still learning from everyone ,each and every day and grateful to one and all in this forum to contribute to my learning.
It took me some time to put perspective to my thought here but I feel I can do it as of now.
No kidding, I just edited a 2.7 year old post of yours, doesn't even sound like you there :)
I know that.. some of my questions are very irrelevant and long lost story like without any connection. The community really helped me learn all along. I have to still go back to them and edit to make sense of it. A few of them very irrelevant questions are garnering a lot of views and I do not want people to get misled by those.
Glad to hear that, please do cite BioStar paper and URL in your paper(s).
and now the post title becomes
Thrown into a Bioinformatics Position & survived :)
this post ignited a light of hope in me...I started bioinformatics last year while I studied plant science even I never worked with linux beforehand. I remember how much I worried the time when I shared this post C: hello i know i must accept that im a stupid after that I changed my account because those days were really sad time for me. I was trying to join a lab but mentor asked me to do a job with BOWTIE2 then he will give me invitation letter for applying for visa...actually biostars helped me to complete my task and get visa! in Berlin I learned a bit more but I hope someday I also share such a post of my victory over bioinformatics as a self learning researcher.
8 months is really fast! Congratulations for putting in the effort and getting a publication out of it :)
Just like you, as a someone who learnt most of bioinformatic stuffs on his own, this post makes me so happy ! Congratulations :)