@@ -122,18 +122,22 @@ __mm10.2bit__ - contains the complete mouse/mm10 genome sequence in the 2bit fil
### lastZ
To align placental mammals, we used previously determined lastz parameters (K = 2400, L = 3000, Y = 9400, H = 2000, and the lastz default scor- ing matrix) that have a sufficient sensitivity to capture orthol- ogous exons
To align placental mammals, we used previously determined lastz parameters (K = 2400, L = 3000, Y = 9400, H = 2000, and the lastz default scoring matrix) that have a sufficient sensitivity to capture orthologous exons
To align placental mammals, we used the lastz alignment parameters K = 2400, L = 3000, Y = 9400, H = 2000 and the lastz default scoring matrix, correspond- ing to parameter set 2 in Table 1. To align non-placental vertebrates, we used K = 2400, L = 3000, Y = 3400, H = 2000 and the HoxD55 scoring matrix. Citation: Increased alignment sensitivity improves the usage of genome alignments for comparative gene annotation. Nucleic Acids Res. 2017;45(14):8369–77.
To align placental mammals, we used the lastz alignment parameters K = 2400, L = 3000, Y = 9400, H = 2000 and the lastz default scoring matrix, correspond- ing to parameter set 2 in Table 1. To align vertebrates, we used K = 2400, L = 3000, Y = 3400, H = 2000 and the HoxD55 scoring matrix. Citation: Increased alignment sensitivity improves the usage of genome alignments for comparative gene annotation. Nucleic Acids Res. 2017;45(14):8369–77.
Create commands for running lastZ for all scaffolds: `lastz.sh`
RepeatFiller [5] is a tool to incorporate newly-detected repeat-overlapping alignments into pairwise alignment chains [4]. Its runtime adds little to the computationally more expensive step of generating chains in pairwise whole-genome alignments. RepeatFiller circumvents the problem that considering all repeat-overlapping alignment seeds during whole genome alignment is computationally not feasible. Therefore, RepeatFiller only aligns local genomic regions that are bounded by colinear aligning blocks, as provided in the chains, which makes it feasible to consider all seeds including those that overlap repetitive regions. RepeatFiller application to mammalian genome alignment chains can add between 22 and 84 Mb of previously-undetected alignments that mostly originate from transposable elements [5]. This helps to comprehensively align repetitive regions and improves the annotation of conserved non-coding elements.
patchChain.perl performs a highly sensitive local pairwise alignment for loci flanked by aligning blocks [1,3]. Given an alignment chain [4], it considers all chains that pass the score and span filters (optional parameters), extracts all the unaligning loci and creates local alignment jobs. After executing these alignment jobs, the newly found and the original local alignments are combined and used to produce a new set of improved chains.
This procedure is recommended for comparisons between species that are separated by >0.75 substitutions per neutral site [1].
`-LRfoldThreshold` = threshold for removing local alignment blocks if the score of the left and right fill of brokenChain. Default is 2.5
`-doPairs` = flag: if set, do test if pairs of chain breaking alignments can be removed
`-LRfoldThresholdPairs` = threshold for removing local alignment blocks if the score of the left and right fill of broken chains (for pairs). Default = 10
`-maxPairDistance` = only consider pairs of chain breaking alignments where the distance between the end of the upstream CBA and the start of the downstream CBA is at most that many bp (default 10000)
`-maxSuspectScore` = threshold for score of suspect subChain. If higher, do not remove suspect.
`-linearGap`=loose
### chainNet
Given a set of alignment chains, chainNet produces alignment nets, which is a hierarchical collection of chains or parts of chains that attempt to capture only orthologous alignments [4]. The original chainNet implementation approximates the score of "sub-nets" (nets that come from a part of a chain and fill a gap in a higher-level net) by the fraction of aligning bases. This can lead to a bias in case the aligning blocks of a chain are not equally distributed. We implemented a new parameter "-rescore" in chainNet that computes the real score of each subnet [2].
can't find database ../data/genomes/mm10.db in hg.conf, should have a default named "db"
```
ERROR with netClass, not sure where to get the "database" from??
### NetFilterNonNested
Before building a multiple alignment from the pairwise alignment nets, it is recommended to remove low-scoring alignment nets that are unlikely to represent real homologies. While the netFilter program [4] removes nested nets in case their parent net does not satisfy the specified score and size criteria, NetFilterNonNested.perl applies a non-nested filtering procedure that considers and filters each net individually [1,3]. This avoids removing nested nets that would satisfy the specified criteria, even if a parent net is removed.
Keeping nets that score higher than 10000 and keeping all nested nets that align to the same locus if they score higher than 3000 for non-placental mammals NetFilterNonNested.perl -doScoreFilter -keepSynNetsWithScore 3000 -keepInvNetsWithScore 3000 -minScore1 10000 ref.query.net.gz > ref.query.filtered.net
FluentDNA looks pretty good for comparing the quality of the alignment. Gives a table of statistics but also provides visualisation tools which looks cool!
https://github.com/josiahseaman/FluentDNA
```
./fluentdna --fasta=data/hg38.fa --chainfile=data/hg38ToPanTro5.over.chain --extrafastas data/panTro5.fa --chromosomes chr19 --outname="Human vs Chimpanzee"
