Cbcb:Pop-Lab:Ted-Report
Jump to navigation
Jump to search
Older Entries
January 15, 2010
Minimus Documentation
Presently, the only relevant Google hit for "minimus" on the first page of results is the sourceforge wiki. The only example on this page is incomplete and appears to be an early draft made during development.
Ideally, it should be easy to find a complete guide with the general format:
- Simple use case:
`toAmos -s path/to/fastaFile.seq -o path/to/fastaFile.afg` `minimus path/to/fastaFile(prefix)`
- Necessary tools for set up (toAmos)
- Other options
- etc
The description found on the Minimus/README page (linked to from the middle of the starting page) is more appropriate, but features use cases that may no longer be common and references another required tool (toAmos) without linking to it or describing how to access it. A description of this tool can be found on Amos File Conversion Utilities page (again, linked to from the starting page), but it is less organized than what I've come to expect from a project page and it is easy to get lost or distracted by the rest of the Amos documentation while trying to peace together the necessary steps for a basic assembly.
Comparative Network Analysis pt. 2
- Meeting with Volker this Friday to discuss how best to apply network alignment to what he's doing
- I'm simultaneously trying to find a way to apply my network alignment technique to predicting genes in metagenomic samples
- I've been trying to find a way to get beyond the restriction that my current program requires genes to be annotated with an EC number. A potentially interesting next step may be to use BioPython to BLAST the sequence of each enzyme annotated in every micro-organism in KEGG against a metagenomic library.
- The results would be stretches of linked reactions that have been annotated in KEGG pathways.
- This method could be applied to contigs just as easily as finished sequences. In a scenario where perhaps there was low coverage, it could be used to identify genes which are probably there but just weren't sampled by showing the presence of the rest pathway. In short, this could finally accomplish what Mihai asked me to work on when I showed up.
- The major theoretical shortcoming of this approach is that it could only identify relatively well characterized pathways.
- The practical shortcoming of this approach will start by obtaining a fairly complete copy of KEGG (which as we've learned is a mess to parse locally and unusably slow to call through the API), and will continue to the computational challenge of such a large scale BLAST operation.
- Ask Bo about this when he gets back. He may have already done this.
- I've been trying to find a way to get beyond the restriction that my current program requires genes to be annotated with an EC number. A potentially interesting next step may be to use BioPython to BLAST the sequence of each enzyme annotated in every micro-organism in KEGG against a metagenomic library.
January 22, 2010
- Met with Dan and Sergey to talk about the Minimus-Bambus pipeline
- Minimus is running fine. I've begun characterizing its run-time behavior (see next week's entry)
- After some tweeking by Sergey, Bambus was able to finish but did not generate a scaffold. We're going to talk about this after the meeting on Monday.
- Sergey had an interesting idea for making a better read simulator:
- Error-free reads are cheap and easy to generate. The problem is with the error model.
- The "best" tool (that we are aware of) which includes error models is MetaSim, but the error models are years out of date and the authors has been historically unreachable. While Mihai has now shown me how to edit the models in a reasonable way from flat files allowing to characterize base substitutions, I'm not convinced it would be faster or easier to write a program that would modify these files than it would be to just write an entirely new program; and given the amount of time I've spent trying to use MetaSim, I'm more than ready to walk away from it. Oh yeah, and MetaSim doesn't work from the command line, so no scripting.
- Sergey has pointed out that most companies will assemble E. coli when they release a new sequencer. Conveniently, there are many high quality assemblies of E. coli available for reference. It might therefore be possible to generate new error models for these sequencers in an automated fashion by mapping the E. coli reads to the available reference genomes, collecting the error frequencies, and then using them to mask synthesized reads.
- I also talked with Mohammad and Mihai about this, who seemed to also think it was a pretty good idea. Mihai has proposed having Sergey or Mohammad add the described error model-generator to his read sampler (written in C) when they have time, but not in preparation of the oral microbiome data.
- Met with James to discuss my work with Volker
- Told him about my meeting with Volker and the paper he wanted me to prepare, more or less by myself. The concepts of the papers are these:
- Most available genomic sequences of mycobacteria are of a very small subset of highly pathogenic organisms.
- Subtractive comparative genomics can be used to identify genes that are potentially responsible for differing phenotypes (such as extreme pathogenicity), but there must be an available genomic sequences for closely related organisms with differing phenotypes.
- Volker has sequenced 2 more non-pathogenic strains of mycobacteria (gastri, and kansasiiW58) with the intention of increasing the effectiveness of these subtractive comparative genomic studies.
- The meat of the paper would be comparing the results of subtractive comparative genomic analysis using all currently available strains in RefSeq, with the results from also using the two novel sequences.
- The other, smaller publishable portion of this project would be a comparison of gastri and kansasiiW58 to each other because they are allegedly thought to be extremely closely related, and yet they have distinct phenotypes (which I've now forgotten).
- James seemed to think this could make an okay paper, and he confirmed that he did not understand that Volker was looking for someone to do all of the analysis, both computational and biological, with Volker only contributing analysis of the analysis after it was all over.
- Ended up also discussing his work on differential abundance in populations of microorganisms.
- I'm going to start working on taking over and expanding Metastats this semester.
- I'm also going to start talking to Bo when he gets back about exactly what he's doing, and how I might be able to include pathway prediction in my expansion of Metastats without stepping on his toes.
- Mihai has given me his approval to focus on this.
- Told him about my meeting with Volker and the paper he wanted me to prepare, more or less by myself. The concepts of the papers are these:
- Met with Mihai to discuss working with Volker
- Explained that rather than looking for someone to do only the complex portions of the computational analysis, Volker was/is looking for someone to do the complete analysis.
- In exchange, Volker is offering first authorship and, if need be, to split the student's funding with their primary PI.
- I think I'm capable of doing this within 3 or 4 months but it would consume my time pretty thoroughly.
- Mihai agreed that this is a reasonable deal, but that I have no personal interest in studying mycobacteria, and it's therefore unwise of me to invest a bunch of time becoming an expert on an organism I have no interest in continuing to study or work with. I've therefore offered Volker to work closely with one of his graduate students who could meet with me every week or two. I would be willing to do all of the computational analysis and explain it to them, but they would have to actually look up potentially interesting genes and relationships I discover and help me keep the analysis biologically interesting and relevant.
- Met with Mihai and Mohammad to discuss our impending huge-ass(embly) problem
- Talked about strategies for iterative assembly as an approach to assembling intractably large data sets. Most have glaring short-comings and complications.
- Discovered Mike Schatz has a map-reduce implementation of an assembler that uses De Bruijn graphs and is better suited to assemblies with high coverage but short read lengths.
January 29, 2010
Minimus Performance Analysis
I'm testing minimus and bambus in preparation of the oral microbiome data, and after spamming several lab members with email, it occurred to me that it would be considerably more considerate to put the information here instead.
| Number of 75bp Reads (in millions): | 1 | 2 | 4 | 8 | 16 | 20 | Model |
|---|---|---|---|---|---|---|---|
| RAM used by the Overlapper (in GB): | 1.2 | 2.4 | 4.5 | 8.7 | 17 | 21.5 | ~1.1 GB * (#Reads in Millions) = (Memory Used) |
| RAM used by the Tigger (in GB): | 3 | 6 | 12 | 25 | 48.4 | (60) | ~3 GB * (#Reads in Millions) = (Memory Used) |
- The 16 million read assembly data is from Walnut, all other numbers are rough averages from both Privet and Walnut.
- Numbers listed in parentheses are predictions made using the listed models.
| Number of 75bp Reads (in millions): | 1 | 2 | 4 | 8 | 16 | 20 | Model |
|---|---|---|---|---|---|---|---|
| Run Time of the Overlapper (in min): | 3 | 9 | 34 | 130 | (576) | 783 | 2.96 * (#Reads in Millions)1.87 = (Run Time in Min) |
| Run Time of the Tigger (in min): | 9 | 66 | 473 | (3,456) | (25,088) | (47,493) | 9.03 * (#Reads in Millions)2.86 = (Run Time in Min) |
- Privet has 2.4GHz Opteron 850 processors and 32GB of RAM. Minimus is not parallelized and therefore only uses a single core.
- Numbers listed in parentheses are predictions made using the listed models.
- The models were generated by plotting the data points in open office and fitting a polynomial trendline. The R2 value for each was 1.
- For reference: There are 1,440 minutes in one day, and 10,080 minutes in one week
| Number of 75bp Reads (in millions): | 1 | 2 | 4 | 8 | 16 | 20 | Model |
|---|---|---|---|---|---|---|---|
| Run Time of the Overlapper (in min): | 2.7 | 8 | 27.5 | 102 | (325) | (481.5) | 2.54 * (#Reads in Millions)1.75 = (Run Time in Min) |
| Run Time of the Tigger (in min): | 14 | 81 | 471.5 | (2,752) | (16,006) | (28,212) | 13.99 * (#Reads in Millions)2.54 = (Run Time in Min) |
- Walnut has 2.8GHz Opteron 875 processors and 64GB of RAM. Minimus is not parallelized and therefore only uses a single core.
- Numbers listed in parentheses are predictions made using the listed models.
- The models were generated by plotting the data points in open office and fitting a polynomial trendline. The R2 value for each was 1.
Other Observations About the Assemblies
- Because of the short read length, every million reads is only 75MB of sequence. This is roughly 10-20x coverage of an average single bacteria. These test sets have reads sampled from roughly 100 bacterial genomic sequences, I would expect the coverage to be on the order of 0.1% on average.
- Unsurprisingly, a cursory glance through the contig files show that each is only comprised of about 2 or 3 reads.
- The n50 analysis for the smaller assemblies shows that only 2-3 reads are being added to each contig on average, leaving both n50's and average lengths just below 150bp.
- Therefore if the complexity of the oral microbiome data is high and/or the contamination of human DNA is extreme (80-95%), the coverage may be extremely low. This may make the use of Mike's assembler impractical, or at least that's how I'm going to keep justifying this testing to myself until someone corrects me.
- Update: Apparently Mike and Dan have talked about this, and somewhere around 75-80bp, the performance of minimus catches up with Mike's de Bruijn graph assembler anyway. I also did not know that Dan's map-reduce minimus was running and would be used to assemble the data alongside Mike's.
- I learned on Feb. 1, 2010 that the 454 error model allows wild variation wrt read length. So these assemblies might not actually be representative of the performance with the illumina data we're expecting on Feb. 20
UMIACS Resources
I just discovered the information listed on the CBCB intranet Resources page is inaccurate and very out of date, so I'm making my own table.
| Machine | Processor | Speed | Cores | RAM |
|---|---|---|---|---|
| Walnut | Dual Core AMD Opteron 8220 | 2.8GHz | 16 | 64GB |
| Privet | AMD Opteron 850 | 2.4GHz | 4 | 32GB |
| Larch | AMD Opteron 850 | 2.4GHz | 4 | 32GB |
| Sycamore | Dual Core AMD Opteron 875 | 1GHz | 8 | 32GB |
| Shagbark | Intel Core 2 Quad | 2.83GHz | 4 | 4GB |
February 5, 2010
Meeting with Volker and Sarada on Feb 3
- Need to teach Sarada how to perform local blast on some sequences they have that aren't yet in genbank
- Trying to set up a meeting with Volker to find out for sure if he wants me to work on this project
Biomarker Assembly
Bo, Mohammad, and I spent a couple hours discussing biomarker assembly today. I'm going to try to efficiently summarize our conclusions, but it might be difficult without an easy way to make images. We eventually decided it would be best to attempt several methods in tandem, due to the severe time constraints. The general approach of each method is to fish out and bin reads through one method or another, and then assemble the reads in each bin using minimus. All sequence identify values will be determined by using BLASTx.
Preliminary Steps
- Gather biomarker consensus amino acid sequences
- Gather amino acid sequences for associated genes from each bacterial genome in refseq
- Cluster amino acid sequences within each biomarker set
Sequence Identity Threshold Determination
There are 31 biomarkers and about 1,000 bacterial genomes in which they occur. This means that there are 31 sets of 1,000 sequences that are all relatively similar to one another. Because of the sequence similarity and the short read length, it's possible that a significant number of reads will map equally well to multiple sequences within each biomarker set. For this reason, it is better to allow a single read to be placed in any bin containing a sequence to which the read mapped above some minimum threshold. This will protect against synthetically lowering the coverage of extremely well conserved regions, and with any luck, incorrectly binned reads will simply not be included in the assembly. There are several ways to approach the determination of this threshold.
- Determine the lowest level of sequence identity between the consensus sequence for each biomarker and any actual protein sequence in that biomarker set. Use that as the minimum threshold for each biomarker set, or use the lowest from any biomarker set as the minimum threshold for all biomarker sets.
- The obvious shortcoming of this approach is that the sequence identity between two homologous gene-length sequences can by much lower than between two homologous read-length sequences.
- Align 75mers to determine the lowest score between any two 75mers in the consensus sequence for each biomarker and the corresponding 75mer in any actual protein sequence in that biomarker set. Use that as the minimum threshold for each biomarker set, or use the lowest from any biomarker set as the minimum threshold for all biomarker sets.
- While this solves the problem with the above approach, it is significantly more complicated and the data is going to be here soon.
- Choose a sequence identity level, or try a few different levels and see which produces the most complete biomarker proteins without creating overly complex graphs.
- While there's no good theoretical justification for this approach, it's probably what we'll do and it will probably work well enough.
Schemes
After making absurdly complicated descriptions of the various approaches which I felt weren't very clear, I used keynote to recreate the diagrams we'd drawn on the white board and then printed them to a PDF. Unfortunately I'm not sure exactly how to embed that in the wiki. So email me at trgibbons@gmail.com if you're reading this and I'll send it to you.
- Marker-wise assembly
- Bin reads that align to any sequence in a given marker set, and/or the consensus sequence for that marker
- Cluster-wise assembly
- Cluster protein sequences
- Bin reads that align to any protein sequence in a given cluster
- Gene-wise assembly
- Bin reads that align to a particular protein sequence
- Marker-wise and cluster-wise binning should be better for assembling novel sequences
- Gene-wise binning should produce higher quality assemblies of markers for known organisms or those that are closely related
February 12, 2010
SNOW!!
February 19, 2010
Met with James to discuss Metastats. I'm going to attempt the following two updates by the end of the semester (I probably incorrectly described them, but I'll work it out later):
- Find a better way to compute the false discovery rate (FDR)
- Currently computed by using the lowest 100 p-values from each sample (look at source code)
- Need to find a more algebraically rigorous way to compute it
- False positive rate for 1000 samples is the p-value (p=0.05 => 50 H_a's will be incorrectly predicted; so if the null hypothesis is thrown out for 100 samples, 50 will be expected to be incorrect)
- James just thinks this sucks and needs to be fixed
- Compute F-tests for features across all samples
- Most requested feature
I spent too much time talking with people about science and not enough time doing it this week...
March 26, 2010
I didn't realize it had been a whole month since I updated. Let's see, I nearly dropped Dr. Song's statistical genomics course, but then I didn't. I did however learn that we don't have a class project. So the Metastats upgrades are going on a backburner for now because ZOMGZ I HAVE TO PRESENT MY PROPOSAL BY THE END OF THIS YEAR!!!
My Thesis Project:
- I'm generally interested in pathways shared between micro-organisms in a community, and also between micro-organisms and their multicellular hosts.
- I'm particularly interested in studying the metabolic pathways shared between micro-organisms in the human gut, both with each other and their human hosts.
- James has created time-series models, and is interested in tackling spacial models with me.
- I would really like to correlate certain metabolic pathways with his modeled relationships.
Volker's Project:
- has taken a big hit this week.
- I'm going to go forward, with Bo's help, using the plan outlined in my project proposal for Mihai's biosequence analysis class:
- Use reciprocal best blast hits to map H37Rv genes to annotated genes in all available virulent and non-virulent strains of mycobacteria
- Use results from gene mapping to identify a core set of tuberculosis genes, as well as a set of predicted virulence genes
- Use a variety of comparison schemes to study the effect on the set of predicted virulence genes of the consideration of different subsets of non-virulent strains
- Use stable virulence prediction to rank genes as virulence targets
Metastats:
- As I mentioned, this is put on hold
- I intend to pick this back up after I'm done with Volker's project, as it could be instrumental to my thesis work
April 2, 2010
More on my Thesis Project:
- I read the most recent (Science, Nov. 2009) paper by Gordon and Knight on their ongoing gut microbiota experiments
- Pretty much every section addressed the potential thesis topics I'd imagined while reading the preceding section. Frustrating, but reaffirming (trying to learn from Mihai on not getting bummed about being scooped).
- Something that seems interesting and useful to me is the do more rigorous statistical analysis to attempt to correlate particular genes and pathways with the time series and spacial data. I will have to work closely with Bo at least at first.
- As a starting point, James has recommended building spacial models similar to his time series models
- James is essentially mentoring me on my project at this point. It's pretty excellent.
July 23, 2010
It's been several months, I don't feel any closer to finding a thesis project, and it's really starting to stress me out. I've finally stopped making excuses for not reading and have been steadily reading about 10 abstracts and 2 papers per week for the last month or two, but it doesn't appear to be nearly enough. I met with Mihai today to talk about it and then foolishly went for a run in the heat of the afternoon, where I decided on a new direction in a state of euphoric delirium.
- Read the book Mihai loaned me within the next week (or so): Microbial Inhabitants of Humans by Michael Wilson
- Mihai says the book is a summery of what was known about the human microbiome 5 years ago. The table of contents for the first chapter is essentially identical to the list of wikipedia pages I've read in the past week, so I'm pretty excited to now have a more thorough, authoritative source.
- Go back to looking for papers describing quorum sensing, especially in organisms known to be present in the human microbiome, either stably or transiently.
- Try not to get too side-tracked reading about biofilms.
- Search for an existing database of quorum sensing genes to use as references to potentially identify novel quorum sensing genes in microbiome WGS data. Consider making one if it's not already available.
- Look for a core metabolome (at this point I think a core microbiome seems unlikely) using metapath (or something similar) in the new HMP data for the gut, oral, and vaginal samples from the 100 reference individuals, as well as other sources like MetaHIT.
- Start with a fast and dirty approach pulling all KO's associated with organisms identified using 16S rDNA sequencing, and then possibly attempt more accurate gene assemblies and annotation from WGS sequencing projects.
- Try to stay focused on a research topic for more than 2 weeks so I don't keep wasting time and effort.
- Don't make a habit of using the wiki like a personal journal...
Possible Research Projects Inspired by Microbial Inhabitants of Humans
- Searching for quorum sensing genes, both known and novel, and any pathways including them in human microbiome data
- Search for and consider making quorum sensing gene DB
- After indexing known quorum sensing genes, search for homologues
- WGS data - Obviously search for homologues directly
- 16S data - Identify organisms and search for homologues in public DBs
- Search for "core metabolome" in pioneer organisms from infant studies